SummaryThe energy hole problem is critical in Wireless Multimedia Sensor Network (WMSN), wherein the nodes closer to the sink node will expire sooner than the outer sub‐regions. Because they transmit their packet and forward the outer sub‐regions packet to the sink. Hence, a hole arises near the sink node after a very short time. Thus, an energy‐aware optimal routing model is introduced in this research using Modified Voronoi‐assisted prioritized double deep Q‐learning (PDDQL). Initially, the sensor nodes are deployed using the Voronoi cell structure that splits the entire region of interest into different Voronoi polygons. However, the Voronoi diagram will not always predict an efficient route because of the nodes at a Voronoi edge or vertex outside the transmission range. To tackle this issue, a new Voronoi diagram is proposed that depends on the maximum horizontal transmission range to remove the portion of Voronoi edges uncovered by the sensor node. After the deployment, a PDDQL algorithm is used to select the Relay Nodes based on their remaining energy. The proposed PDDQL determines an alternate route to transmit the packet if an energy hole problem has occurred in WMSN. The analysis of a proposed routing protocol based on the assessment measures like packet delivery ratio, average residual energy, number of alive nodes, and average end‐to‐end delay accomplished the value of 198.33 (ms), 2.05 (J), 100, and 0.85, respectively.