BackgroundPrevious studies have reported that the prevalence of depression and depressive symptoms was significantly higher than that before the COVID-19 pandemic. This study aimed to explore the prevalence of depressive symptoms and evaluate the importance of influencing factors through Back Propagation Neural Network (BPNN). MethodsData were sourced from the psychology and behavior investigation of Chinese residents (PBICR). A total of 21,916 individuals in China were included in the current study. Multiple logistic regression was applied to preliminarily identify potential risk factors for depressive symptoms. BPNN was used to explore the order of contributing factors of depressive symptoms. ResultsThe prevalence of depressive symptoms among the general population during the COVID-19 pandemic was 57.57 %. The top five important variables were determined based on the BPNN rank of importance: subjective sleep quality (100.00 %), loneliness (77.30 %), subjective well-being (67.90 %), stress (65.00 %), problematic internet use (51.20 %). ConclusionsThe prevalence of depressive symptoms in the general population was high during the COVID-19 pandemic. The BPNN model established has significant preventive and clinical meaning to identify depressive symptoms lay theoretical foundation for individualized and targeted psychological intervention in the future.