<p>In view of the uneven resource allocation and insufficient efficiency of the traditional power data management system based on deep learning, edge computing is used to build the power data management system framework. The deep deterministic policy gradient is introduced to address the resource allocation optimization of the power data management system. The solution network is used to optimize the deep deterministic policy gradient, and a deep reinforcement learning intelligent power data management system driven by edge computing is designed. The results showed that when the iteration was 400, the value function of the Actor network in the designed network was 357.6, significantly higher than that of other networks. The loss of the Critic network in the designed network was 0.06, significantly lower than that of other networks, and the optimization effect was good. When the number of users was 3, 4, 5, and 6, the energy efficiency range was [1.1, 0.7], [1.5, 1.0], [2.1, 1.7], and [2.6, 2.1], respectively, which was significantly lower than that of other methods. The designed method reduces the operating cost of the system. When the network parameter was 0.5, the average delay was 8.7&#xa0;s. When the network parameter was 0.8, the average delay was 27.4&#xa0;s, and the response speed was fast. The above results exhibit the superiority of the designed network and effectively promote the further intelligent development of power data management.</p>