To address the phenomenon of temporary communication blindness resulted from fixed period beacon exchange in mobile wireless sensor networks, an adaptive beacon exchange algorithm is proposed. The key idea is that work node calculates variable beacon period according to characteristic value relative to up node. idle node calculates variable beacon period according to characteristic value relative to its all neighbors.The threshold probability can be adjusted to meet the performance requirement of networks. Forwarding node removes the next hop from neighbors table if its overtime to wait for the feedback beacon. The simulation shows that the adaptive beacon exchange algorithm can acquire high reach rate for eliminating the phenomenon of temporary communication blindness, especially in work-node-sparse sceneries, with low consumption. So the algorithm is scalable and applicable to large-scale mobile wireless sensor networks.