Theoretical and experimental aspects of oxygen isotope fractionation in carbonate minerals are critically examined based on a direct comparison of fractionation factors for carbonate-water systems. The results show good agreement between theory and experiment in most cases. In particular, theoretical fractionation factors calculated by the statistico-mechanics method and the increment method are in good agreement with the experimental values for dolomite, siderite, witherite, strontianite and cerussite. These agreements provide corroboration that the two entirely independent approaches of calculation are generally capable of producing thermodynamic equilibrium fractionation factors for the most carbonate-water systems. In particular, the merit of the increment method relative to the statistico-mechanics method is evident for crystalline minerals because it enables the systematic and accurate predictions of oxygen isotope fractionation factors for different structures and compositions of crystalline minerals based only on their crystal chemistry. Thus, the increment method has no limitations to the calculations as commonly encountered by the statistico-mechanics method. Nevertheless, complexity in oxygen isotope fractionations between calcite and the other carbonate minerals can be caused by the oxygen isotope inheritance during polymorphic transformation from aragonite to calcite and by the isotope salt effect on the other carbonates in the presence of aqueous fluids. Therefore, caution must be exercised when interpreting possible disagreements between theory and experiment because of the kinetic effects. In the extreme case, equilibrium oxygen isotope fractionation factors for single carbonate-water systems could be either underestimated or overestimated by one of the theoretical and experimental methods.