In order to obtain the optimal structural parameters of a centrifugal compressor casing treatment slot, the optimization of centrifugal compressor casing slot parameter was carried out. A BP neural network prediction model was established based on the existing simulation data, and a multi-objective optimization of the slot structural parameters was carried out by using the genetic algorithm (NSGA-Ⅱ). The results show that the optimized structural parameters are the slot width of 4 mm and the slot center position of 1.5 mm from the middle section of air guide wheel near the leading edge of impeller. According to the simulation analysis, the slot treatment structure corresponding to the optimized value is relatively close to the leading edge of impeller, which makes the low-speed region relatively forward and improves the flow situation of main channel so as to shift the surge boundary obviously to the small flow rate at high speed, the stable working range of compressor widens.