The invention discloses an internal control re-check type associated control-based hybrid class rapid prototyping method. According to the method, a deep learning strategy is utilized to construct aninitial control aggregation to ensure that control heat source bodies are uniformly distributed in a control space; and re-check type associated control is adopted to perform local control on an optimal control heat source body so as to find an optimal solution. Since simplex deterministic local control is only performed on the elite control heat source body in an evolution process, so that the number of the times of the calculation of an algorithm fitness function can be greatly decreased; the deep learning strategy can make the control heat source bodies uniformly distributed in the controlspace, thereby laying a foundation for global control; the Cauchy mutation operator is adopted, so that the method of present invention can be prevented from being trapped in local optimum; and finally, the global optimal solution is obtained in the shortest time.