Abstract： Objective:Based on the BP neural network model，to establish a multi-dimensional fitting algorithm for nitrogen supersaturation safety factor for the calculation of diving decompression scheme.Methods:Based on the Haldane decompression model，and adopting the nitrogen supersaturation safety factor in the 12-60 m Air Diving Standard Decompression Table，a neural network fitting model of nitrogen supersaturation safety factor was established and evaluated by analyzing the mapping relationship between diving parameters and nitrogen supersaturation safety factor.Results:The BP-based neural network fitting model of the nitrogen supersaturation safety factor was established，and the minimum value of MAPE on the off-bottom test set was 0.438 6%，and R2 was 0.998 7. While the minimum value of MAPE at the stop station test set was 0.528 9%，and R2 was 0.995 5. Conclusion:The nitrogen supersaturation safety factor obtained by the BP neural network fitting method is close to the one used in actual diving decompression schemes，which can be used for formulating real-time diving decompression schemes.