There is an increasing demand for the wheelchairs which could make decisions by acquiring information from the surrounding, for the wheelchair users all around the globe. Such type of wheelchairs is called intelligent wheelchairs and are using different types of high level controlling algorithms to control itself. Using EMG based controller for high level decision making has become a reliable and efficient method due to the high signal to noise ratio of EMG signals and the ability of them to bring out the users’ intentions successfully. Usually, those EMG based controllers are incorporated with another modality to achieve the desired tasks through the intelligent wheelchairs. An EMG based high level controller which can be stand-alone without the aid of other modalities is rarely found in the literature. Even the available such type of controllers is mostly designed for the users with full/partial upper limb functions. However, such types of controllers cannot be used by the wheelchair users with trans-humeral amputation/ trans- radial amputation due to the absence of the required muscles. This paper proposes an EMG based high level controlling algorithm for the navigation input generation and to detect the intention for using a robot manipulator for reach to grasp task, for an intelligent wheelchair with a robotic manipulator. The proposed controller has been designed for wheelchair users with trans- humeral amputation/ trans- radial amputation. But it can be used by any type of wheelchair user with neck functions and full/ partial functions of both upper limbs up to the humerus. Experiments have been done with the healthy human subjects to validate the efficacy of the proposed high level controller.