This chapter deals with the estimation of pairwise genetic distances using various nucleotide substitution models. Simple proportional differences (p-distance) are contrasted with corrected measures that account for multiple hits, saturation and varying substitution rates. Examples demonstrate how p-distance underestimates divergence in distantly related taxa. The chapter then explores different substitution models, their assumptions and applications, stressing the importance of choosing appropriate models for robust phylogenetic inference. Model selection is positioned as an essential prerequisite for likelihood and Bayesian approaches discussed later, highlighting the interplay between evolutionary processes and computational methods.