Flushing time is an integrative parameter that can represent water exchange capacity and sensitivity to pollution threats. As bays formed where tributaries enter the Three Gorges Reservoir of China have experienced frequent algal blooms over the past decade, we examine the spatially variable flushing time of a typical tributary bay (Xiangxi Bay) by numerical tracer experiments. First, the tracer concentration reduction could be approximated well by a double exponential decay curve, and the local flushing time of Xiangxi Bay is determined using key flushing coefficients. Second, a sensitivity analysis is performed to examine the influence of upstream inflow, temperature difference, wind, water level of the reservoir, and daily water level fluctuation on the spatial variation in local flushing time. Finally, according to local flushing time values and the sensitivity analysis results, the bay can be generally characterized by three zones: riverine, transitional, and mainstream-influenced zones. In particular, the local flushing times in the riverine zone are mainly affected by the upstream inflow. The difference in temperature between the reservoir mainstream and Xiangxi Bay is the main forcing in the transitional and mainstream-influenced zones. This study is the first to investigate different driving factors for flushing time in a typical reservoir tributary bay. The findings provide insights on the transport processes in such water bodies, suggesting the possibility of using the longitudinal zonation of flushing time for reservoir management.