This study evaluates credit scoring models using statistical regression techniques in consumer finance, aiming to enhance risk assessment accuracy and financial inclusion. A quantitative research design was employed, analyzing secondary data from 2020 to 2024. Logistic regression, multiple linear regression, and support vector machines were applied to assess their predictive power. Findings indicate that logistic regression achieved an 85.3% accuracy rate (AUC = 0.78), while advanced models like support vector machines and neural networks yielded 92.3% and 91.1% accuracy, respectively. Key predictors of creditworthiness included credit history length (-0.74 correlation with default rate), income level (-0.52), and age (-0.26). The overall correlation coefficient between predictor variables and credit risk was -0.62, confirming a strong inverse relationship. The study concludes that while logistic regression remains a robust credit scoring tool, hybrid models incorporating machine learning techniques offer superior predictive performance. It recommends integrating alternative credit data, regular model updates, and bias-mitigation strategies to enhance credit risk assessment and ensure equitable access to finance.