This study develops an intelligent energy predictor using a Bidirectional long short-term memory (BiLSTM) neural network model for predicting energy demand response within a smart home. To implement this proposed model, the real-time data from five different appliances such as air conditioner, TV, computer, fan, and lamp were monitored via smart meters within a Home Energy Management System (HEMS). In addition to the smart meters, an extra meter has been installed to monitor overall energy consumption data. This real-time collected data setup enables the collection of precise and comprehensive energy consumption data, which in turn supports data analysis and confirms the system’s effectiveness before its large-scale deployment. The data normalization and feature selection plays a crucial role in the effectiveness of the proposed model. Various performance metrics are used to validate the proposed model and compare with LSTM model. The proposed Bi-LSTM method achieved superior performance over the LSTM model by reducing errors interms of performance metrics such as MAE, RMSE, NRMSE, MSE, and improving R2 of individual appliances in a smart home. The results highlight the model’s potential in smart-home energy management and suggest future work in integrating contextual data and enhancing real-world applications.