This thesis reviews, applies and evaluates the ability of credit scoring models to quantify the credit risk of personal loans in Libya. After discussing the various theoretical and practical issues involved in credit risk modeling, the thesis first provides an overview of the main characteristics of the evolving consumer banking system in Libya, with a particular emphasis on the procedures currently applied by Libyan commercial banks when making decisions granting personal credit. Attention is especially given to the limitations of the credit assessment process currently employed by Libyan banks and its impact on successful credit risk management. This is followed by a comprehensive international survey of the extant empirical work on credit scoring models. This provides a comprehensive methodological framework for the parametric and nonparametric modeling of consumer credit risk, including variable specification. In the empirical part of the thesis, confidentialised personal loan records provided by Libya‘s Al-Wahada Bank are randomly sampled and used along with discriminant analysis, logistic regression and survival analysis to construct and evaluate the competing credit scoring models. While the results identify a small common set of independent variables that could be considered as the most important variables affecting the probability of the default of personal loans, including the loan amount, the borrower‘s income, whether the guarantor‘s salary is ongoing, whether the borrower guarantees another borrower, and whether the guarantor guarantees another borrower, the findings also indicate that logistic regression performs significantly better than either discriminant or survival analysis in consumer credit scoring models in the chosen context.