This paper provides a comparative analysis of three machine learning models used to predict wind power generation, which are Support Vector Machine (SVM), Nearest Neighbours (KNN), and Random Forest (RF). The primary objective is to evaluate the effectiveness of the model for short-term operational planning to support transmission system operators (TSOs) in grid management and power system stability decisions. The analysis utilizes real wind farm SCADA system data from the Wind Time Series Dataset. The results show that SVM significantly outperforms KNN and Random Forest, achieving RMSE of 2.288 MW, R2 score of 0.986 and MAPE of 4.090%, demonstrating superior accuracy and generalization for operational forecasting. Data preprocessing challenges including missing values and skewness were addressed through KNN imputation and logarithmic transformation. A comprehensive dashboard tool shows the reallife application of a suggested real-time wind power forecast application to TSO operations and micro-grid management. The study recommended TSOs to maximise wind power forecasting systems, to help improve grid stability and the integration of renewable energy effectively. This contributes to enhanced grid reliability and efficient integration of renewable energy.