Owing to the pressure of on-time availability of products to the customers, the order from suppliers should come on time. To predict the lead time accurately, machine learning models are required but those models work for the organization order data. There is need of inclusion of external event data like natural disaster, social and political events etc. for better prediction of lead time. Therefore, the present paper provides an approach of supplier order lead time prediction. For lead time prediction, a unique approach has been deployed which is amalgamation of various models like segmentation, prediction and simulation. The segmentation is done by implementing unsupervised learning and prediction is done by implementing four different techniques like Linear Regression, Regression Tree, XGBoost and Random Forest while Monte Carlo Simulation is used for simulation. The proposed approach has been tested on the numerical example which is based on simulated data. The prediction accuracy increases if external event effect is added. The prediction is done using four different machine learning techniques while the simulation is done using Monte-Carlo Simulation.