The traditional soundscape assessment method assesses people’s perception of a selected acoustic environment using questionnaires or interviews. However, fewer studies have been done to investigate other methods for methodological triangulation. This research aims to propose an interactive approach for soundscape assessment research based on participatory design and apply this method to study the relationship between audio-visual elements and soundscape perceptions in residential areas. The experiment was conducted in a virtual reality (VR) environment. Thirty-two participants were invited to interactively create the audio-visual environment, which they thought was the most eventful or vibrant in outdoor and indoor contexts. The result indicated that the descriptor “eventfulness” could be related to “vibrancy” since participants selected similar audio-visual elements for these two descriptors except for building and greenery. Moreover, participants preferred to model open visual environments and more greenery in vibrant scenes, while they preferred semi-closed buildings and less greenery in eventful scenes. Besides, the result revealed that people in the indoor context might be more sensitive to traffic sound than in the outdoor context. Those results were consistent with those of traditional assessment methods, which means this method may be applied for a methodological triangulation in soundscape evaluation.