AC algorithm of blind sources has been applied to the identification of engine noise sources. However, the testing platform based on the blind separation theory is difficult to actualize some experiments and obtain data in the real environment, or to obtain these requires high cost. Therefore, this paper proposes a new approach to construct an experimental testing platform-Sinmulink Model so as to identify engine surface noise sources. This model is applied to the test of the experimental data. These experiments demonstrate that Simulink model is efficient to identify engine noise source and the testing data accord with that made by the former platform. It is obvious that Simulink model, to some degree, can be used to replace engine surface noise sources testing platform.