The achievement of wireless communication technology of mobile devices has been witnessed, which produces a large number of trajectory data of mobile users. Processing and analyzing these trajectories could obtain users' movement patterns and behavior rules, leading to provide better location-based services such as point-of-interest recommendation and location prediction. However, enormous volumes of GPS trajectory with high frequency will pose challenges in storage, transmission and computation. Recently, with the rise of social networking sites, more and more users tend to share their geographic locations in real time, thus forming check-in sequences. Hence, this paper proposes a model called SSTLP for predicting next location from trajectory based on spatiotemporal sequence. Firstly, construct location transition probability model by capturing the change of locations in historical trajectory. Secondly, compute distance possibilities of locations by the combination of normal distribution and cosine similarity and then the next location could be figured out. Experiments on real-world data set demonstrate that the proposed model outperforms traditional prediction algorithms.