<p>This paper proposes a new control cooperation scheme which enables a system of uncertain arm manipulators not only to track the desired trajectory but also to handle the unknown factors. On the one hand, a new disturbance compensation inverse dynamic centralised controller is first proposed to handle the mission of tracking the desired point. On the other hand, the responsibility of coping with the unknown factors is assigned to a disturbance estimator that possesses online learning capabilities through a Radial Basis Function neural network. To achieve these objectives, the mathematical models of arm manipulators are linearised by combining all known models into a short Euler–Lagrange system, and all uncertain terms are lumped into a unique vector. Then, the disturbance estimator will compensate for uncertain terms. The inverse dynamic centralised controller and disturbance estimator are also designed and operated together, ensuring that the system’s states track the prescribed reference model. Therefore, the control scheme in this paper is named lumped disturbances estimation-based inverse dynamic cooperation control or adaptive inverse dynamic controller for short. Besides offering benefits for application to the uncertain manipulators, this new control scheme is also simple to perform in real projects when controlling task is handled by a single centralised controller. The mathematical expressions are proven sufficiently and the numerical simulation is verified through the Matlab/Simulink platform.</p>