The sensor network that contains sensor nodes equipped with cameras, microphones and other sensors producing multimedia content is known as wireless multimedia sensor network (WMSN). In WMSN, battery energy and network lifetime are the real requirements of research for the transmission of multimedia information needs more energy. This research paper proposes a soft computing-based approach for optimizing the cluster head selection process to decide upon the optimal path of transmission, and SGP approach is used for cluster formation. A three-layer ANN model trained using back propagation has been proposed in this paper, and comparative analysis is performed among existing techniques such as LEACH, Q-LEACH, C-LEACH and SGP algorithms. It is found from the observations that proposed neural hybrid approach exhibits superior performance while exhibiting marginal improvement over SGP due to the training process where the error is minimized.