AbstractStream discharge is a key hydrological factor for water supply planning, wetland loss investigation, ecological service assessment, and climate change impact estimation. Conceptually, stream discharge is expected to be highly and positively related to precipitation. In reality, however, such a relationship may be weaker because precipitation characteristics are affected by local climate of watersheds. For many watersheds around the world, a vast amount of precipitation data are readily available but the stream discharge data are very limited or unavailable. It would be time‐saving and cost‐effective to predict stream discharge based on precipitation data. Unfortunately, this task is very difficult to achieve using the traditional methods. Although the copula method is able to establish a good relationship (or a good dependence structure) between discharges and precipitations, this relationship does not include the time series process, and thus is impractical for applications. Therefore, a hybrid of copula prediciton and time series computation was developed (with detailed procedures) here to estimate stream discharge based on precipitation data. The method was validated using the measured daily discharges with the good statistical measures, that is, the Kandell’s τ (0.42–0.44), normalized root mean square error (2.19–2.28 m3/s), and R2 (0.66–0.84). This study suggests that the hybrid method is a useful tool to predict stream discharges based on precipitation data.