The nighttime light (NTL) remote-sensing data have been widely applied in several applications for analyzing the urbanization process. The relationship between NTL intensity and human activity becomes a solid foundation for the applications using NTL data. However, there is no research, so far, revealing how the human activity seasonality could impact the seasonal change of NTL intensity. In this paper, a comparative analysis, box plot, and random forest algorithm were applied to NTL remote-sensing data and points of interest (POIs) data within Shanghai, China. The results show that in spring and autumn, the NTL is much brighter than that in summer and winter, especially within high human activity density area. The NTL intensity can be partly (approximately 40%) explained as the joint effects of the five POI categories. By analyzing the contributions of each POI category to NTL intensity, we found that the National Polar-Orbiting Partnership-Visible Infrared Imaging Radiometer Suite (NPP-VIIRS) could be used to dig more information about gross domestic product (GDP) and traffic-based applications with consideration of NTL seasonality.