In this paper, we present a neurocontroller via adaptive learning rates (ALRs) for stable path tracking of mobile robots. The self recurrent wavelet neural networks (SRWNNs) are employed as two neurocontrollers for the control of the mobile robot. Since the SRWNN combines the advantages such as the multi-resolution of the wavelet neural network and the information storage of the recurrent neural network, it can easily cope with the unexpected change of the system. Specially, the ALR algorithm in the gradient-descent method is extended for the multi-input multi-output system and is applied to train the parameters of the SRWNN controllers. The ALRs are derived from the discrete Lyapunov stability theorem, which are used to guarantee the stable path tracking of mobile robots. Finally, through computer simulations, we demonstrate the effectiveness and stability of the proposed controller.