The article presents organization of a knowledge base required for implementation of the intelligent control of the movement of an android robot hand with account of the positions of the exclusion zones. The proposed structure of the knowledge base characterizes the previous experience of the synthesis of the hand movements by the velocity vector with account of the known obstacles, and sets the properties intrinsic to them. The knowledge base consists of four components. The first one specifies the images, which represent a projection of the contours of the working areas of the mechanism, taking into account the exclusion zones. The second one stores configurations with the maximal values of the parameters, which define the solid angle, used to calculate the most optimal initial set position with respect to an object of manipulation. The third one cmoanxtains configurations of the movement, from which a deadlock situation results. The fourth one sets the maximal values of kimax parameters. The fourth component of the knowledge is necessary to reduce the time of calculation of the vector of the generalized velocities, when the calculated configuration crosses the exclusion zone. The algorithm of the virtual motion control of a robot android organized on the knowledge base allows us to implement the methods for evaluation of the current situation and make respectively one or another logical choice. The results of the computational experiments show a reduction of time for calculation of the test tasks associated with setting and removal of the manipulation objects on the shelves and racks with the use of the developed knowledge base.