Smart grid operations involve electricity load management through renewable energy generation. Taking into account the diverse spread of the power consumers and non-reliable nature of renewable energy resources, an accurate energy consumption forecasting model would help create an efficient and demand management plan of operation. Developing accurate and high-speed short-term load forecasting is essential for the development of the energy market and smart grids. In this paper, Random Vector Functional Link Network-based deep neural model was implemented for the prediction of total electric load consumed by the hour. The standard dataset used here comprises the various sources of power generation and the amount of power generated by each of these sources is used. Based on the energy features extracted by the deep neural model, a correlation between the amount of energy generated through various sources and the total consumption of energy is established. The load prediction model has performed with a Mean Squared Error loss of0.115 and was evaluated using other metrics such as Mean Absolute Error of prediction and predictive R-squared score.