In this current era 5G technology used in system wireless access method for communication. In this research paper proposes and designs develop a lightweight NN for the recognition of facial expressions. Where a visual system that can be implemented into devices with a limited computational capacity to perform face expression classification while also reducing the number of parameters used in the classification process. Enhancing the accuracy of the findings including encapsulating the models into a consumable service that can be used by other collaborating systems. The genetic resource creation frameworks for the face expressional approach may be planned to meet the needs of various security models including criminal exposures, governmental confidential security problems, and so on. this model will clearly opposing effect on detection and classification. and obtains the good detection results by identifying images inside the dataset, which proves that the model designed in this paper is suitable for multi- classification of facial expressions. In general, comprehended a visual system that can be integrated on devices with low computing power to attain result classification and reduce a large number of parameters. After comparing with the algo in recent years, the accuracy of our algo will higher than theirs, and it has achieved good detection results in images inside the dataset after the experimental results.