A nonlinear adaptive control algorithm for a multi-task redundant manipulator is developed in this paper. The method considers the parametric uncertainties in the system and defines a proper filtered error signal according to the allocated priority. Based on this error analysis, the asymptotic stability and convergence of the tracking error, both for the main task as well as the sub-tasks are shown using Lyapunov approach. In order to extend the algorithm for the case of orientation control in the operational task, quaternion feedback has been exploited, and the stability is shown. The results of the paper are verified in several simulations on 4DoF planar arm as well as 7DoF KUKA lightweight robot arm.