A numerical approach is proposed for minimum‐torque trajectory planning (MTTP) of kinematically redundant manipulators between two joint configurations. Different from the traditional pseudoinverse control scheme, the first step of the proposed approach is to use an iterative procedure to transform the MTTP problem into a nonlinear programming (NLP) problem. Then an approach to generate a series of feasible solutions for the NLP problem is proposed. Since a sufficient number of feasible solutions can be easily generated, the optimization process of the NLP problem can be started from different points to search for all the local minima. To show the validity of the proposed approach, a planar three‐link robotic manipulator is included in the simulation example.