Abstract: The domestic zinc smelting industry exhibits a relatively low level of intelligent automation. Replacing manual labour with slag removal robots in zinc melting furnaces can mitigate safety risks and enhance equipment intelligence. However, factors such as the complex furnace environment and the extended length of the slag rake make precise trajectory tracking challenging, with tracking accuracy directly impacting slag removal quality. Consequently, trajectory tracking control research is conducted for multi-arm slag removal robots, designing PID and sliding mode control strategies. This aims to enhance slag removal performance, providing theoretical reference and practical value for slag removal robot research.