 This paper proposes a deep-ensemble neural network for predicting hourly power consumption. The model was trained on 13 data sets spanning 2004-2018, with two columns for date, time, year and energy expenditure. The data was normalized using Minmax scalar before being fed into the model. The model was evaluated using various statistical metrics such as RMSE, MABE, R2, MBE and MABE. The results showed that the proposed model outperformed other models, demonstrating its accuracy in predicting power consumption. This article was authored by Mohamed Irfan, Ahmad Shaf, Tariq Ali and others.