<p>Under the cooperative–competitive framework, a mixed Pareto differential game-based approximate optimal tracking control for modular robot modeled by the Newton–Euler iteration approach under uncertain disturbance is investigated in this paper. Under the mixed Pareto differential game, the tracking issue of the modular robot under disturbance is transformed into a Pareto differential game among modules of robot and zero-sum game between the robot controller as well as disturbance. The coupled Hamilton–Jacobi–Bellman–Isaacs (HJBI) function is solved by the critic neural network (NN) which approximated the cost function; thus, the optimal control pair, namely each module’s optimal controller and worst disturbance is obtained. Experiment platform is verified by the preponderance of the proposed method.</p>