This paper presents a joint selection criterion for optimal trajectory planning using dynamic programming along a specified geometric path subject to torque and velocity constraints. A single non-stationary joint is normally considered for optimization purpose. In case of more than one non-stationary joints, the optimization by selecting any joint at random may end up with a non-optimal solution. This problem raises the importance of the proposed joint selection criterion. A novel joint selection criterion based on the geometric path information is presented for dealing with such cases having more than one non-stationary joints. In contrast of existing approaches to solve the optimization problem by considering each joint separately, the proposed criterion saves computational time, efforts and thus there is no need to solve the optimization problem multiple times with respect to each non-stationary joint to get the optimal solution. The proposed criterion can be used to optimize a given path with respect to travelling time, energy consumption or any other arbitrary form of cost function. The validity of the proposed selection criterion is tested on a parallel DELTA robot. The obtained results are compared with the optimal results obtained by other techniques which confirm the applicability of the proposed scheme.