Landscape design research is important for cultivating spatial intelligence in landscape architecture. This study explores GIS (geographic information systems) as a tool for landscape design research - investigating landscape designs to understand them as architectonic compositions (architectonic plan analysis). The concept ‘composition’ refers to a conceivable arrangement, an architectural expression of a mental construct that is legible and open to interpretation. Landscape architectonic compositions and their representations embody a great wealth of design knowledge as objects of our material culture and reflect the possible treatment of the ground, space, image and program as a characteristic coherence. By exploring landscape architectonic compositions with GIS, design researchers can acquire design knowledge that can be used in the creation and refinement of a design. The research aims to identify and illustrate the potential role of GIS as a tool in landscape design research, so as to provide insight into the possibilities and limitations of using GIS in this capacity. The critical, information-oriented case of Stourhead landscape garden (Wiltshire, UK), an example of a designed landscape that covers the scope and remit of landscape architecture design, forms the heart of the study. The exploration of Stourhead by means of GIS can be understood as a plausibility probe. Here the case study is considered a form of ‘quasi-experiment’, testing the hypothesis and generating a learning process that constitutes a prerequisite for advanced understanding, while using an adjusted version of the framework for landscape design analysis by Steenbergen and Reh (2003). This is a theoretically informed analytical method based on the formal interpretation of the landscape architectonic composition addressing four landscape architectonic categories: the basic, the spatial, the symbolic and the programmatic form. This study includes new aspects to be analysed, such as the visible form and the shape of the walk, and serves as the basis for the landscape architectonic analysis in which GIS is used as the primary analytical tool. GIS-based design research has the possibility to cultivate spatial intelligence in landscape architecture through three fields of operation: GIS-based modelling: description of existing and future landscape architectonic compositions in digital form; GIS-based analysis: exploration, analysis and synthesis of landscape architectonic compositions in order to reveal latent architectonic relationships and principles, while utilizing the processing capacities and possibilities of computers for ex-ante and ex-post simulation and evaluation; GIS-based visual representation: representation of (virtual) landscape architectonic compositions in space and time, in order to retrieve and communicate information and knowledge of the landscape design. Though there are limitations, this study exemplifies that GIS is a powerful instrument to acquire knowledge from landscape architectonic compositions. The study points out that the application of GIS in landscape design research can be seen as an extension of the fundamental cycle of observation, visual representation, analysis and interpretation in the process of knowledge acquisition, with alternative visualisations and digital landscape models as important means for this process. Using the calculating power of computers, combined with inventive modelling, analysis and visualisation concepts in an interactive process, opened up possibilities to reveal new information and knowledge about the basic, spatial, symbolic and programmatic form of Stourhead. GIS extended the design researchers’ perception via measurement, simulation and experimentation, and at the same time offered alternative ways of understanding the landscape architectonic composition. This gave rise to the possibility of exploring new elements in the framework of landscape design research, such as the visible form and kinaesthetic aspects, analysing the composition from eyelevel perspective. Moreover, the case study showcases that GIS has the potential to measure phenomena that are often subject to intuitive and experimental design, combining general scientific knowledge of, for instance, visual perception and way-finding, with the examination of site-specific design applications. GIS also enabled one to understand the landscape architectonic composition of Stourhead as a product of time, via the analysis of its development through reconstruction and evaluation of several crucial time-slice snapshots. The study illustrates that GIS can be regarded an external cognitive tool that facilitates and mediates in design knowledge acquisition. GIS facilitates in the sense that it can address the ‘same types of design-knowledge’ regarding the basic, spatial, symbolic and programmatic form, but in a more precise, systematic, transparent, and quantified manner. GIS mediates in the sense that it influences what and how aspects of the composition can be understood and therefore enables design researchers to generate ‘new types of design-knowledge’ by advanced spatial analysis and the possibility of linking or integrating other information layers, fields of science and data sources. The research contributes to the development and distribution of knowledge of GIS-applications in landscape architecture in two ways: (1) by ‘following’ the discipline and developing aspects of it, and (2) by setting in motion fundamental developments in the field, providing alternative readings of landscape architecture designs.