One essential aspect of supply chain management in the pharmaceutical industry is selecting the optimal shipping method for delivering crucial medications to patients efficiently and on time. We aim to automate the selection of optimal shipping modes for different medications based on product characteristics, price, and delivery urgency. A comprehensive pharmaceutical supply chain dataset is taken, encompassing various features such as shipping costs, drug expiry dates, and delivery dates. Data cleaning and feature engineering are performed to drop irrelevant features, transform the data, and extract meaningful insights from the data set of 10,000 instances. This paper evaluates the performance of five ML algorithms: LightGBM, XGBoost, Random Forest, Decision Trees, and CatBoost. ML models are also evaluated based on their training time and complexity. LightGBM demonstrates superior performance metrics such as accuracy and f1-score and achieves the highest accuracy of 96% in predicting shipment modes of pharmaceutical drugs. Our findings show the potential of gradient boosting methods for optimizing shipment mode. Accurately predicting the best shipping option can reduce transport costs, enhance delivery timeliness, and increase patient satisfaction.