The utilization of non-renewable energy resources necessitates the power sector's adoption of alternative energy sources, including photovoltaic and wind power generation systems. This academic investigation utilizes two machine learning methodologies, in particular, the study utilizes the random forest and support vector machine algorithms. to conduct its analyses. predict the velocity of the wind in the Diyala governorate of Iraq for the subsequent time interval. This is achieved solely by utilizing historical monthly time series data as input predictors. The three performance metrics employed encompass the coefficient of assurance (R2), root cruel square mistake (RMSE), and cruel outright blunder (MAE). The findings demonstrate that utilizing a lag of 12 months in the time series data (the maximum lag duration tested) as input predictors leads to the most accurate predictions in terms of performance. However, the prediction performance of the two algorithms used was almost similar (RF's RMSE, MAE, and R2 were 0.237, 0.180, and 0.836, while for SVM were 0.223, 0.171, and 0.856). The capacity to anticipate wind speed constitutes a paramount advantage to Iraq, given its current predicament in the electric power industry, and this has the potential to enable stakeholders to forecast oversupply or undersupply and implement pre-emptive measures.