Street trees are highly visible components of the urban forest, providing ecosystem services to city residents, but tree mortality can result in a loss of these essential benefits. Here, we focus on how street tree survival in Philadelphia, Pennsylvania (US) has been influenced by 1) tree traits and attributes, 2) the local biophysical environment, and 3) socioeconomic context. We used a longitudinal monitoring approach with a 6-year time interval and 103 randomly-located, fixed-length permanent plots, containing a range of tree size classes, to assess street tree survival rates, predictors of survival, and the balance of tree losses and gains. Street trees surveyed in this study experienced an annual mortality rate of 3.68% and followed a Type III mortality curve, with the highest annual mortality rates in the smallest size class. On the surveyed plots, there was a net decrease of live trees over the study period (2657 to 2542; 4.33% decline). Multivariate logistic regression revealed significant associations between street tree survival and tree crown vigor, species traits (expected mature stature and drought tolerance), site type, land use type, % impervious surface in a 15 m buffer, parcel-level demolition events, % of residents in poverty, % unoccupied households, and % renter-occupied households. Additionally, the final model included a multiplicative interaction term between crown vigor and tree size with separation in pattern of small (DBH < 7.6 cm) from larger (DBH ≥ 7.6 cm) trees. Our findings suggest that targeted management interventions may enhance overall street tree survival, such as focusing on small trees, as well as trees in sidewalk cut-outs, commercial or industrial areas, or near demolition activity. Longitudinal monitoring studies such as ours provide valuable information about how urban forest systems change over time, with assessments that are valuable for management by municipal foresters and other street tree stewards.