The accuracy of steering control of an underwater glider is very important for ocean target detection. Current steering control of the underwater glider (UG) mainly uses proportional-integral-derivative (PID) controller. However, to ensure that the underwater glider moves in accordance, PID controller parameters need to be repeatedly set and adjusted, which makes it difficult to meet the requirements for fast and accurate control. To solve the problem, a parametric self-tuning PID control method based on the radial basis function (RBF) neural network was proposed. Firstly, the dynamic model of the underwater glider in the horizontal plane was established. Then, the RBF neural network structure was constructed, and the iterative formulas of neural network parameters and PID parameters were given by the gradient descent method. Simulation results show that compared with the conventional PID controller, this controller has shorter setting time, higher precision, and the parameters of the controller can be quickly self-tuned. It provides a reference for the design of the underwater glider steering controller in the future.