Abstract: During the welding process, the fluctuation of the contact force between the welding gun of the welding robot and the double toothed roller crusher will be transmitted to each joint through the robotic arm, forming a load torque disturbance, resulting in dynamic changes in the damping ratio and natural frequency of the control system, making it difficult to achieve accurate tracking control. Therefore, a welding seam tracking control method for intelligent welding robot of double toothed roller crusher is proposed. By adaptively adjusting the region of interest to capture weld seam features, a kinematic model of an intelligent welding robot is constructed, the motor angular velocity adjustment is calculated, and the position and posture of the welding gun are adjusted in real time through coordinate transformation theory. Introducing Lyapunov stability constraints, dynamically adjusting fuzzy PID parameters based on welding gun position, adaptively adjusting weld deviation compensation values, stabilizing damping ratio and natural frequency, and achieving weld seam tracking control. From the experiment, it can be seen that the proposed method is basically consistent with the actual weld trajectory, with small errors, and the anti-interference ability of the method is more advantageous.