Zebrafish is a useful animal model for studying human diseases such as muscle disorders. However, manual monitoring of fish motion is time-consuming and prone to subjective variations. In this paper, an automatic fish motion analytics framework is proposed. The proposed framework could be exploited to help validate zebrafish models of transgenic zebrafish that express human genes carrying mutations which lead to muscle disorders, thus affecting their ability to swim normally. To differentiate between wild-type (normal) and transgenic zebrafish, the proposed framework consists of two approaches to exploit discriminative spatial–temporal kinematic features which are extracted to represent zebrafish movements. First, the proposed approach studies precise quantitative measurements of motor movement abnormalities using a camera with the capability to record videos with high frames rates (up to 1,000 frames per second). This differs from previous works, which only tracked each fish as a single point over time. Second, the proposed approach studies multi-view spatial–temporal swimming trajectories. This differs from previous works which typically only considered single-view analysis of fish swimming trajectories. The proposed motion features are then incorporated into a supervised classification approach to identify abnormal fish movements. Experimental results have shown that the proposed approach is capable of differentiating between wild-type and transgenic zebrafish, thus helping to validate the zebrafish models.