This study aims to design a PID (Proportional-Integral-Derivative) controller that can tune nonlinear systems effectively. The PID controller uses LSTM (Long short-term memory), an RNN (Recurrent Neural Network) algorithm to enhance its performance, using time series data to predict PID parameters, allowing the controller to adjust nonlinear systems accurately and efficiently. Furthermore, the effects are discussed through the peak amplitude and transient time of various graphs, as well as the characteristics of the inverted pendulum. In this study, it is found that the LSTM deep learning model can produce better and more significant results when the mass of the pendulum is greater than that of the pendulum cart, while the Ziegler Nichols method doesn’t perform as precisely. As for the inverted pendulum system, the amplitude graph fails to converge when the KP value is less than 1. The graph becomes increasingly divergent when the KP value decreases. In conclusion, the LSTM deep learning model can be applied to nonlinear systems to increase its stability and to find suitable PID parameter values for the system promptly.