To better construct the economic management system, the author proposes a method based on nonlinear algebraic equations. Using BDS statistics as the test method of data model correctness, it is proved that BDS statistics can be used to test the correctness of a large class of data models. Based on the actual economic data, the validity of the BDS statistical test method and its theory is verified. The experimental results show that the BDS statistics obtained are 5.77 and 7.11. It can be seen that BDS statistics reject AR (2) model. This implies that some part of the sequence has a large variance, which is derived from a different system, and therefore leads to the rejection of the null hypothesis. It is proved that BDS statistics can be used to test the correctness of a large class of data models.