The adequate cost estimation of mill plants plays a crucial role in the success of feasibility studies of mining projects. Grinding is one of the most important operations in mineral processing plants and assumes a substantial share of the total milling costs. The objective of this work was to develop a set of cost functions for major grinding mill equipment. These cost models were developed using two relatively different techniques: uni-variate regression (UVR) as well as multivariate regression (MVR) based on principal component analysis (PCA). The first is appropriate for the quick estimation of costs in the early stages of project evaluation, while the second method can be helpful in the feasibility study stage. The explanatory variable in UVR was power (P), while in MVR the power and some other variables depending on the type of mill were used. The PCA technique was employed in order to omit the correlation between the independent variables in the multivariate regression. Furthermore, the scale-up factor for all mills has been calculated. The result of the evaluation of the models showed that the mean absolute error rates were less than 9.84% and 11.36% on average for the capital and operating costs of the uni-variate model, and 5.82% and 4.9% for the multivariate model, respectively. (C) 2013 Elsevier Ltd. All rights reserved.