The airline industry has grown tremendously in the last two decades. This growth is accompanied with challenges, one of which is flight delay. Flight delay is a serious setback that results in loss of billions of dollars each year. The aviation industry is highly dependent on punctual flight operations, and delays can incur substantial costs, inconvenience passengers, and affect airline reputation. To mitigate the impact of flight delays, this research presents an innovative approach to analyse and predict flight arrival delays using a hybrid machine learning technique. This approach combines the strengths of two machine learning algorithms, in two stages, while stage one classifies the delay, stage two estimates the delay time in minutes enhancing prediction accuracy and robustness. The model was evaluated with classification and regression metrics. After a comprehensive data preprocessing and modelling, a Recall score of 94%, R-Square score of 85.07%, MAE of 20.72 and RMSE of 15.92 minutes was obtained. The resulting model offers accurate predictions, real-time adaptability, and valuable insights into delay patterns. This paper contributes to improved operational efficiency, enhanced passenger experiences, and cost savings for airlines and airports operators.