Abstract: Void fraction is an important parameter in gas-liquid two-phase flow, which is closely related to the flow pattern, pressure drop, and heat transfer of the flow field. Since flow boiling is widely encountered in various industrial fields, especially in the cores of pressured water reactors (PWRs) and boiling water reactors (BWRs), accurate predicting of the void fraction is beneficial for controlling the reactor reactivity. There are many methods for the prediction of void fraction, among which the slip-ratio-based method is the most popular one. However, since it is quite difficult to measure the variations of slip ratio in experiments accurately and the relevant theoretical research is still lack, there exist significant differences of prediction accuracy between different slip-ratio-based void fraction models. Therefore, it is necessary to conduct a comprehensive evaluation of existing slip-ratio-based void fraction models through multiple evaluation indicators. On this basis, 8 slip ratio models were collected in this paper, which considered the influence of physical properties and vapor quality. Then, an experimental database for void fraction was established, containing 380 sets of experimental data of air-water two-phase flow. The overall performance of these slip-ratio-based void fraction models was comprehensively evaluated both based on the error analysis and the grey correlation analysis. The evaluation indicators mainly include the mean absolute error, root mean square error, error frequency distribution, and grey correlation degree, among which the grey correlation degree can reflect the similarity between the predicted values and experimental results quantitatively. The results indicate that for air-water two-phase flow, the Chisholm correlation performs best within the full range of void fraction (0<α<1). Due to the influence of flow patterns, there exist large differences in the performance of various slip ratio correlations within different void fraction ranges. In the low-range of void fraction (0<α<0.3), bubbly flow is the main flow pattern. The gas phase content is relatively low, and the velocities of two phases may be unevenly distributed on the cross-section of the channel, resulting in the low accuracy of the slip-ratio-based models. Thus, slip-ratio-based models are not recommended in this range to predict the void fraction. However, with the increase of the gas content, both gas and liquid phases exist in the form of continuous phases, which significantly improves the performance of these models. As a result, Lockhart&Martinelli correlation performs best in the moderate range of void fraction (0.3<α<0.7). The Chisholm correlation performs best in the high range of void fraction (0.7<α<1).