<p>In this paper, we address the problem of learning time-varying graph models from spatiotemporal signals by jointly capturing spatial and temporal dependencies. Existing graph learning methods are predominantly centralized, resulting in high computational costs for large-scale data, while distributed methods that consider only spatial correlations face limitations in handling dynamic signals. To overcome these challenges, a novel distributed graph learning algorithm is proposed, which leverages temporal correlations to efficiently learn evolving graph structures. Experimental results on synthetic and real-world datasets demonstrate that the proposed algorithm achieves competitive accuracy compared to centralized methods and outperforms existing distributed approaches.</p>