This paper presents an experimental and statistical study on the kerf width, used instead of the width of the cut in abrasive waterjet (AWJ) cutting. Pre-dimensioned granitic rocks were sampled for the experimentations designed by using Taguchi orthogonal arrays. The effects of the AWJ operating variables on the kerf width were studied and the rock properties were correlated with the kerf widths. Additionally, predictive models for the kerf widths were developed using multi-variable regression analysis and the developed models were verified through some statistical tests. The results demonstrated that the standoff distance and the traverse speed have significant effects on the kerf widths. The results also showed that water absorption, unit weight, microhardness, the maximum grain size of rock-forming minerals, and mean grain size of the rock have significant correlations with the kerf widths of the tested rocks. Furthermore, the modeling results revealed that the predictive models derived from rock properties, can be successfully used as a practical guideline.