The human perception of environments is a complex mechanism, influenced not only by visual and auditory stimuli and their perceptual outcomes, but also by personal characteristics, site types, and other factors that affect people's restorative experiences. However, there is little research that treats it as a complete systemic phenomenon to reveal its integrated impact. We surveyed 1161 visitors in seven urban parks in Hangzhou, China, and collected information on audio-visual stimuli. Using structural equation modeling and multi-group analysis, we initially examined the comprehensive impact of acoustic factors, non-acoustic factors, and individual characteristic variables on perceived restorativeness. Subsequently, we studied the effects of different park types on the constructed multivariate comprehensive model. The results confirm that the presence of different types of sound sources within landscape features can impact the auditory affective quality dimension, which in turn affects visitors' perceived restorative experiences. In addition, the reported mediating pathways revealed that visual elements related to nature, by providing high-quality visual landscapes, enhance the pleasantness of soundscapes and yield greater restorative benefits in park environments. A new finding of the study was that, although the eventfulness of the soundscape negatively affects perceived restorativeness, the presence of high-quality visual landscapes in the environment can compensate for this deficiency. Further multi-group analysis results show that landscape features and visual perceptual quality indirectly influence perceived restorativeness through the auditory affective quality dimension, resulting in varied correlations across two types of park. Moreover, the pleasantness of soundscapes consistently has a significant effect on perceived restorativeness in both types of parks; however, in parks oriented towards activity, the eventfulness dimension of soundscapes does not significantly affect perceived restorativeness. Lastly, the influence of landscape features on perceived restorativeness, mediated by types of sound sources, is both direct and indirect, leading to diverse correlations across different park environments. These detailed findings and guidelines for audio-visual interaction design will aid in the planning and development of urban soundscapes.