Mini-parks are becoming a popular form of outdoor recreational space in densely populated areas, largely because their small size makes site selection easier than for ordinary parks. However, existing studies on mini-parks are limited because most of them rely on data collected through traditional surveys, which are severely restricted by space and time. In this study, we utilised Tencent user density data – a type of space–time synchronous data with high spatial resolution – to trace mini-park visitation in the main city of Yancheng, China, and we integrated data about land use, points of interest, transportation, demographics and housing prices to measure the parks’ surrounding features. We investigated how factors relating to the parks’ spatial and internal attributes, surrounding physical features and surrounding socio-economic features affected the number of park visits during the week versus during the weekend by establishing a series of multiple linear regression models. The results showed that higher resident population, more surrounding public toilets and larger open site promoted mini-park visits while distance to the city centre, surrounding large parks and main roads discouraged mini-park visits. This study also found that the effects of weekend visitation factors were more complex than those of weekday visitation factors. These findings can help urban green space planners and decision makers to efficiently allocate mini-parks to areas where they will be most effective.