Abstract: Kinematics modeling of the developed welding mobile robot was presented and a real time seam tracking algorithm based on fuzzy neural network was proposed.At first, the kinematics behavior of mobile robot body and cross sliders action to the welding torch were investigated by the Denavit-Hartenberg Homogeneous transformation method, and a full kinematics model was established. And then a seam tracking controller based on fuzzy-gaussian neural network (FGNN) was described by applying a Gaussian function as an activation function, taking lateral slider position and heading angle of the robot as input signals, and the adjusted angle for welding torch as output, a specialized learning architecture was used so that membership function ould be tuned in real time by applying the backpropagation algorithm of FGNN controller.The experiment results show that the proposed controller has excellent tracing accuracy (within±0.5 mm), and can satisfy the requirement of practical welding project.