Human activity discovery aims to cluster the activities performed by humans, without any prior information of what defines each activity. Most methods presented in human activity recognition are supervised, where there are labeled inputs to train the system. In reality, it is difficult to label activities data because of its huge volume and the variety of activities performed by humans. In this paper, an unsupervised approach is proposed to perform human activity discovery in 3D skeleton sequences. First, important frames are selected based on kinetic energy. Next, the displacement of joints, statistical displacements, angles, and orientation features are extracted to represent the activities information. Since not all extracted features have useful information, the dimension of features is reduced using PCA. Most human activity discovery proposed are not fully unsupervised. They use pre-segmented videos before categorizing activities. To deal with this, we have used sliding time window to segment the time series of activities with some overlapping. Then, activities are discovered by a hybrid particle swarm optimization with Gaussian mutation algorithm to provide diverse solutions. Finally, k-means is applied to the outcome centroids from each iteration of the PSO to overcome the slow convergence rate of PSO. Experiments on three datasets have been presented and the results show the proposed method has superior performance in discovering activities in compared to the other state-of-the-art methods and has increased accuracy of at least 4 % on average.