In order to obtain the best technological parameters for laser cutting of 304 stainless steel with a thickness of 3 mm, a laser cutting machine with a power of 3 kW was used for laser cutting experiment. A four-factor and three-level orthogonal experiment scheme was designed to study the effects of laser power, cutting speed, oxygen pressure and defocusing amount on laser cutting quality. Through poor analysis to explore the weight of each factor on laser cutting quality, and then by the orthogonal experiment data, a BP neural network prediction model was set up and use method of combining the BP neural network and genetic algorithm to find the best parameter combination. Comprehensive optimization to obtain the optimal process parameters of laser cutting quality, the laser power is 750 W, the cutting speed is 2 000 mm/min, oxygen pressure is 0.5 MPa, and the coke is +2 mm. The test results show that the optimized specimen has smooth cutting surface, small roughness and slit width, and meets the expected requirements.