<p>Accurate dynamic model is crucial for broad robotic applications including precise control, fast collision detection and adaptive trajectory planning. This paper focuses on the dynamic modeling and parameter identification of a heavy load baggage handling robot by considering nonlinear friction. Firstly, a basic dynamic model is established based on the Newton–Euler algorithm and a nonlinear friction model, i.e., the Stribeck model. Particularly, nonlinear effects of the joint tilt and end payload on friction torque are considered by fitting the corresponding functions of friction torque. Secondly, a two-step procedure for parameters identification is proposed based on the weighted least squares algorithm: (1) friction models of the joints are identified with a uniform reciprocating excitation trajectory. (2) Basic inertial parameters of the robot are identified with Fourier series function excitation trajectory. Finally, a real baggage handling robot system consisted of an ABB IRB-6700 robot and a pallet-type end effector is used to verify the proposed friction models and dynamic parameter identification method. The results show that the accuracy of the identified joint torques can be effectively enhanced.</p>