Abstract: To improve the utilization of ground test data in predicting the dynamic environment of aircraft during flight, a response mapping prediction method based on a probabilistic neural network (PNN) was proposed, taking into account the nonlinear dynamic characteristics of complex aerospace structures and the uncertainties contained in response data. Firstly, the theoretical basis of the response mapping prediction method under distributed load was given, indicating that the establishment of the mapping relationship was independent of the number of load sources. Then, the necessity of introducing PNN to establish mapping relationship was analyzed, and the specific details of using the PNN method for response prediction were emphasized. Finally, the proposed method was verified by a noise test of an instrument cabin. The results showed that the PNN method has good prediction accuracy in different load levels and full-frequency bands. In addition, the nonlinear dynamic characteristics of the instrument cabin under noise excitation were analyzed. The results showed that the PNN method has better prediction accuracy than the matrix mapping method. This paper extends the deterministic mapping prediction method to the probabilistic mapping prediction method, which improves the credibility and engineering practicality of the mapping prediction method.