The study reported in this paper deals with path planning in three-dimensional space using network optimization for robotic manipulators working in the presence of obstacles and workspace singularities. To execute the algorithm, the robot design parameters, the geometry and location of the obstacles, and the initial and goal positions of the desired task are required. As a first step, the manipulator workspace is discretized and points inside forbidden regions formed by obstacles and singularities are excluded. An ellipsoidal searchspace is then selected as a part of the workspace to make the network enumeration and path synthesis more efficient. Based on an allowable deviation angle, path segments axe created to form the connectivity network. A path which is optimal with respect to the manipulator kinematic and dynamic properties is generated as a sequence of intermediate points connecting the initial and goal states using Dijkstra’s minimum cost search technique. A computer program has been developed to implement this methodology for three-axis manipulators, and results of the application of this algorithm to some industrial robots are presented.