The continued increase in national energy demand pushes oil and petroleum price prediction efforts for the net oil-importing Philippines to ensure adequate supply. These prices are commonly modeled by data-driven Machine Learning (ML) methods to encompass their extrinsic and volatile nature. However, recent studies have found that new features specific to COVID-19 and the Russia-Ukraine geopolitical conflict have significantly contributed to overseas oil price ML prediction. This work investigated the impact of these factors on seven pump prices in Manila, Philippines, from December 2019 until July 2022. Three ML regression models were chosen to extend the existing literature and ensure price model accuracy: Multiple Linear Regression (MLR), Support Vector Regression (SVR), and Random Forest Regression (RFR). Features were listed based on related literature and underwent data preprocessing using p-value testing and Principal Component Analysis. Models were then trained, tested, and optimized using nested splits and hyperparameter tuning. Mean Absolute Percent Error (MAPE) was used to evaluate accuracy. Generated models had MAPE values within the range 3.13 % - 12.67 %, which is within the range of MAPE values in oil and petroleum price prediction literature, 0.131 % - 19.2 %. MLR and SVR models generally exhibited the highest accuracy for each pump price. This study proves that period-specific features may be used for local pump price modeling. Future works may explore other ML models and geographic location effects and investigate newly identified period-specific features.