The wine industry is a global and multifaceted sector that encompasses the production, distribution, and sale of wine. The profitability of the wine industry is influenced by various factors, including consumer preferences and production costs. Wines with higher alcohol levels often feel fuller, have better aging potential, and intensify the perception of certain aromas, contributing to a more expressive and aromatic wine. Proper prediction of alcohol level based on the chemical properties that can be controlled can significantly reduce wine production costs and save resources on research and wine trials. Machine learning (ML) models can analyze chemical properties of grapes to predict alcohol levels in the final product enabling winemakers to adjust production processes accordingly and achieve the desired alcohol content. This pipeline applies major regression ML algorithms to guide winemakers in adjusting production processes to enhance desired white wine characteristics. Results show that Light and Extreme Gradient Boosting machines, and Random Forest methods have very high and almost similar accuracy in predicting alcohol levels. However, Categorical Boost(CatBoost) Regressor presents the highest prediction accuracy values according to all regression performance metrics. Here is the link to the video presentation of the code: https://www.loom.com/share/c264a36862a6411a8599b3e3b0d126ba?sid=0d4b2be4-dd11-435e-afe0-e176458a23df