This paper presents an algorithm to generate the single support phase trajectory in each step for exoskeleton users with low strength in their lower body. This method generates the trajectory of the single support phase according to the patient\\\'s body parameters and enables the patients to voluntarily change their walking speed. In this method, the body parameters of the exoskeleton user and the information of the double support phase are taken as input. A multi-layer perceptron neural network, human kinematic model, and cubic spline interpolation are used to generate the single support phase trajectory for the hip and knee joints. To evaluate the performance of this algorithm, a healthy person walked on the ground at three different speeds, and the walking data was recorded. Then the knee and hip joints trajectory of the single support phase for one step was generated by the proposed algorithm. Then the generated Trajectories were compared with the actual trajectories. The experimental results reveal that the proposed algorithm is able to generate a smooth personalized trajectory according to the preferred walking speed.

Keywords