Over the last decade, regeneration of derelict and underused sites with varying degrees of contamination (also known as Brownfield sites) has gained popularity as a sustainable land use strategy. However, redevelopment of contaminated fields is a complex and multidimensional problem that entails many risks and uncertainties. The objective of this thesis is to construct, calibrate and validate a risk assessment model that can assist investors and decision-makers in evaluating and classifying brownfield sites to two categories : suitable for redevelopment / not suitable for redevelopment. The three-step model building process is adopted from the methodology of credit risk modeling used in banks and credit rating agencies. The proposed models utilize two machine learning algorithms, namely Classification And Regression Trees (CART), and Random Forest algorithms. The first part of the thesis provides a point of reference in browfield regeneration risk modeling and describes the current research gaps in this field. The following chapter describes the credit risk model building methodology. Finally, Chapter 4 describes the implementation of risk model building methodology in the field of brownfield risk modeling using programming language R. Appendix A includes the commented Rcode for interested readers and can serve as a guideline in implementing the Classification And Regression Tree, and Random Forest algorithms in various fields of study.