This project aims to develop a website that functions as a virtual laboratory for studying geography topics that is suitable for use, as well as testing its effectiveness on cognitive learning outcomes and spatial abilities. The research uses the Borg and Gall development model, which is then summarized and divided into three main activities: introduction, development, and testing. Subject selection was based on the cluster random sampling technique. The subjects consisted of Class Learning outcome data was collected using questionnaires, interviews, pretests and posttest, analyzed using descriptive statistical tests and Mancova inferential statistics. Research Results: 1) the development of a geographic virtual laboratory is suitable for use according to material experts and received the "very good" category, while according to media experts it received the "good" category; 2) students gave a strong response to the implementation of the virtual geography laboratory; 3) there is a significant difference between initial abilities and cognitive learning outcomes and spatial abilities; and 4) there are no significant differences between two schools that implement virtual geography laboratories with cognitive learning outcomes and spatial abilities.