Utilizing control chart pattern recognition techniques allows for determination of whether the current production process is performing abnormally, thereby enabling quality analysis and facilitating worker identification of anomalies in a timely manner. The GA-BP algorithm is a control chart pattern recognition algorithm. However, it exhibits certain shortcomings such as slow convergence speed and a fixed maximum number of iterations. The fixed maximum number of iterations implies that the algorithm may terminate before convergence, resulting in decreased accuracy. In response to these problems, this paper proposes an improved GA-BP algorithm, which employs adaptive learning rate and adaptive iteration number methods. Specifically, the algorithm first uses a GA algorithm to obtain the optimal initial weight values of the BP network. Based on this, the algorithm updates the weight values using adaptive learning rate methods and updates the maximum number of iterations using adaptive iteration number methods. These improvements significantly increase the convergence speed of the algorithm while also avoiding early termination of iterations. Experimental results demonstrate that the proposed algorithm significantly improves classification accuracy and reduces training time.