The rapid advances in networking technologies aspect foremost crisis in terms of the system reliability, efficiency, sustainability and cost in concert with the wireless sensor communication network. This communication platform integrated with Smart Grid system to transverse us the information about the grid system failure, fault detection, retrieval, security, monitoring and management of the energy bases with assurance. At present, cognitive radio networks (CR) is vital among the wireless technology for a smart grid communication system with less hindrance and proficient bandwidth.This paper aims at introducing a contemporary data traffic scheduling framework in a CR-SG communication network based on an advanced sailfish optimized Deep Convolutional neural network (SF-DCNN) for an effectual Smart grid communication system. Here, perhaps a strengthened DCNN is unwrapped to schedule the traffic data and so elements using the selfish optimizer algorithm to achieve less latency with maximum throughput. The results of the analysis demonstrate that the proposed approach on the SG system will executes substantially better scheduling than all other perceived models, suggesting that the prototype proposed is well suited to standardized the communication in a SG system with less interference.