In rapidly growing cities, on-street parking spaces consistently serve as a supplement to urban parking supply. However, prolonged curb occupancy remains a prevalent issue on roadways with limited parking management, leading to reduced traffic flow and safety concerns in non-motorized areas. To support the formulation of fine-grained management strategies based on on-street parking duration (OPD), we investigate the distribution of OPD across various roadway segments and explore the factors influencing it at different scales. Real data are obtained from a newly built-up area of Xi'an, China, and the spatial effects of factors influencing OPD are modeled using a multiscale geographically weighted regression (MGWR) method. Results show that the number of parking events, the number of parking spaces, employment buildings, the parking space location, and the distance to adjacent bus stops spatially significantly affect the OPD on both weekdays and weekends. The lowest parking fee level is positively associated with weekday OPD. The estimated coefficients of the variables for on-street parking lots are spatially heterogeneous, which is discussed with the substantial parking policy implications. This study can inform enhanced parking management policies in cities through spatial effects analysis. •On-street parking duration (OPD) is investigated in newly built-up urban areas.•The multiscale GWR is utilized to capture the spatial effects of OPD influences.•Policy implications for parking management are given with further discussion.