This paper addresses the problem of detecting a dim moving target in a sequence of hyperspectral images using snapshot techniques. In these applications, the target is low contrast relative to background clutter, which makes designing a detection algorithm with a low false alarm rate and high detection probability a challenging task. In this paper, we propose a novel spatial-temporal anomaly approach to solve this problem. First, a spatial anomaly map is calculated by the proposed dual-window-based rapid RXD (r-DWRXD). Then, a temporal anomaly detector is proposed to effectively suppress background clutter, which has the advantage of not requiring accurate frame registration. In addition, we address motion consistency by introducing a trajectory history map, which is calculated using a simplified variance filter (VF) to detect the presence of targets from the temporal profile of each pixel in multiple spatial anomaly maps. After fusing the spatial anomaly map, temporal anomaly map and trajectory history map, the target of interest can be easily detected from the background. The proposed approach is applied to a test dataset of airborne targets in a sky background. Experimental results confirm that the proposed approach can achieve better detection performance than other state-of-the-art approaches.