Wind speed (WS) forecasting is essential for optimizing wind power generation. The increasing adoption of renewable energy has made the ability to accurately predict WS increasingly important. Therefore there is need to compare predictive techniques based on different activation functions. For this ReLU, Leaky ReLU, Swish, and Exponential Linear Unit (ELU) are used as activation function for Artificial Neural Network (ANN) model. The accuracy is compared and it is found that ELU activation function provide better result with mean square error, root mean square error, mean absolute error and correlation value as 0.3998, 0.6323, 0.4832 and 80.2% respectively.