Parkinson’s disease is a gradually advancing neurological condition marked by tremors, stiffness, and difficulties in movement, stemming from the decline of cells responsible for producing dopamine in the brain. The goal is to enhance prediction accuracy by analyzing protein biomarkers in the cerebrospinal fluid. Advanced machine learning models such as Random Forest, Gradient Boosting, Support Vector Regression (SVR), and Classification and Regression Tree (CART) are employed in this study to analyze biomarkers and other features in predicting the progression of Parkinson’s Disease (PD). These models are then assessed using metrics such as Mean Average Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R-squared (R2) scores, and the results are interpreted using Explainable AI for better clarity. The average mean square error for models trained without other UPDRS scores was 5.37 for Random Forest (RF), 5.53 for Gradient Boosting (GB), 5.73 for Support Vector Regression (SVR), and 5.59 for Classification and Regression Trees (CART). In contrast, models trained with other UPDRS scores exhibited lower average mean square errors, with values of 3.53 for RF, 3.57 for GB, 4.80 for SVR, and 3.77 for CART. Apolipoprotein C-III influences UPDRS 1, while Retinol-binding protein and Phosphatidylcholine-sterol acyltransferase affect UPDRS 2. Immunoglobulin heavy constant alpha 1 plays a role in UPDRS 3, and Alpha-1B-glycoprotein is linked to UPDRS 4. The findings show the promising capabilities of these methods in Perkinson’s diagnosis, offering new insights into disease progression.