In cold regions, freeze-thaw cycles are one of the most important agents affecting on the mechanical properties of stones and consequently their durability. Uniaxial compressive strength, Brazilian tensile strength and P-Wave velocity are among important mechanical properties in assessing the durability of the stones used in cold regions. As regards of determination these properties during the freeze-thaw cycles experiment is very time consuming and cumbersome, statistical relationships can be used to estimate the mechanical properties. In this study, a multivariate statistical model is presented for 15 samples of travertine, which can ​ estimate uniaxial compressive strength, Brazilian tensile strength and P-Wave velocity in each cycle of the freeze-thaw experiment. In this model, mechanical properties after freeze-thaw cycles were considered to be the dependent variable that dependent on the variables of the initial mechanical properties, water absorption and the number of freeze-thaw cycle. The statistical tests results and also data from other researchers show that model presented is reasonably accurate in estimating uniaxial compressive strength, Brazilian tensile strength, P-Wave velocity and consequently their durability of the travertines in each cycle of the freeze-thaw experiment.