Soundscape is an essential component of urban forest landscapes, acoustic indices can be effectively used to monitor biodiversity, but whether they can be used for soundscape perception assessments needs to be further explored. In this study, soundscape recordings were collected in Beijing Eastern Suburban Forest Park, and acoustic indices were used to explore the relationship between the acoustic environment and soundscape perception, as well as the possible effects of temporal changes. To achieve this, audio recordings collected in spring and summer were divided, and a total of 90 audio segments were extracted from three time periods—morning, afternoon, and evening—to calculate the acoustic index and complete a questionnaire survey. The urban forest soundscape was evaluated according to the eight perceptual attribute quality indicators of ISO 12913, and generalized linear models were constructed to quantify the relationships between the acoustic indices and perception. The results showed that the temporal variation of the soundscape influenced the subjective evaluation, with the highest overall evaluation relating to the morning soundscape. The combination of acoustic indices explained the soundscape pleasantness (R² = 0.58) better than the soundscape eventfulness (R² = 0.54), demonstrating the utility of these indices in soundscape assessment. Linking acoustic indices to human perception generates innovative ideas and theoretical support for soundscape enhancement, contributing to a more pleasant acoustic environment and maximizing the social value of urban forests.