Since the surface alloying process has multi-input and multi-output, non-linear coupling and time varying dynamic characteristics, it is very difficult to establish an accurate process model for designing model-based controller. Hence an adaptive neural network controller is developed in this paper to tackle the variation and disturbance of the laser alloying control system. The proposed control strategy is based on a neural network structure combined with sliding-mode control scheme. An adaptive rule is derived based on the reaching condition of a specified sliding surface to on-line adjust the weights of radial basis functions neural network (RBFNN). It has on-line learning ability in response to the system's non-linear and time-varying behaviors. Two adaptive neural network controllers are designed for regulating the laser power control and traverse speed, respectively, to achieve the desired temperature distribution on the alloying surface. The control performance is evaluated based on numerical simulation results. Different disturbances are introduced into the system output variables to investigate the robustness of the proposed controller with respect to practical process variation.