In this paper, an optimal adaptive sliding mode controller of an Omni-Directional Mobile Robot (ODMR) is proposed using harmony search algorithm. First, kinematic model of the robot is derived and then, governing equations of dynamic model have been obtained using linear and angular momentum equilibrium. Since the derived model is not an exact definition of the system, it includes some uncertainties. To compensate them, a tracking control method has been offered. The proposed controller consists of an approximately known inverse dynamic model output as the model-based part of the controller, an estimated uncertainty term to compensate for the un-modeled dynamics, external disturbances, and time-varying parameters to enhance closed-loop stability and account for the estimation error of the uncertainties. In order to compare the results of the proposed controller, an optimal feedback linearization and sliding mode controllers are designed and then, a cost function has been defined by combining the variation rate of control signal and the integral error index. This cost function has been minimized using harmony search algorithm, resulting in optimum control parameters. Finally, the performance of the designed controller in different conditions, such as in presence of disturbance and system parameter variation has been simulated and discussed.