The transportation sector is a significant contributor to global CO2 emissions, driving the urgent need for innovative approaches to reduce environmental impact. This study explores the application of machine learning (ML) models, specifically Random Forest and XGBoost, to predict vehicular CO2 emissions using openly available data from the Canadian Government Open Data Portal. Key vehicle attributes—such as engine size, fuel consumption, and transmission type—were analyzed to identify the primary drivers of emissions. Through exploratory data analysis and feature selection, eight critical features with the highest correlation to CO2 emissions were identified and used in model training. Both ML models demonstrated high predictive accuracy, with R2 values of 0.98, explaining 98% of the variance in emissions. SHapley Additive exPlanations (SHAP) values were employed to interpret the model predictions, highlighting fuel consumption metrics, engine size, and the number of cylinders as the most influential factors. These findings underscore the potential of ML to provide actionable insights into the factors driving vehicular emissions, informing policies that could enhance fuel efficiency and engine technology to mitigate climate impact. Future research could extend this framework by integrating real-time data and evaluating emerging technologies, such as electric vehicles, to further support climate change mitigation.