Predictive policing has been used as a governance tool in China for a decade, but as a governance object, it has only recently entered the research field. There are three major risks associated with predictive policing: prediction accuracy, service normativity, and social adaptability. Given the inherent uncertainties of predictions and the limitations of interventions, the aspiration to stopping crime before it begins remains a challenge. The future aim should be to elevate prediction accuracy and, within ethically and legally acceptable limits, convert predictive insights into tangible governance practices.