Energy management is a vital tool for reducing significant supply-side deficits and increasing the efficiency of power generation. The present energy system standard emphasizes lowering the total cost of power without limiting consumption by opting to lower electricity use during peak hours. The previous problem necessitates the development and growth of a flexible and mobile technology that meets the needs of a wide variety of customers while preserving the general energy balance. In order to replace a partial load decrease in a controlled manner, smart energy management systems are designed, according to the preferences of the user, for the situation of a full power loss in a particular region. Smart Energy Management Systems incorporate cost-optimization methods based on human satisfaction with sense input features and time of utilization. In addition to developing an Internet of Things (IoT) for data storage and analytics, reliable LoRa connectivity for residential area networks is also developed. The proposed method is named as LoRa_bidirectional gated recurrent neural network (LoRa_ BiGNN) model which achieves 0.11 and 0.13 of MAE, 0.21 and 0.23 of RMSE, 0.34 and 0.23 of MAPE for heating and cooling loads.