Numerical simulation methods are usually time-consuming when predicting the impact characteristics of water droplets on the airfoil. In order to quickly and accurately calculate the droplet mass caught near the airfoil under icing conditions， a rapid prediction method based on proper orthogonal decomposition （POD） and the surrogate model is proposed. First， the optimal Latin hypercube sampling method is employed for the continuous maximum icing conditions in FAR 25 Appendix C， and the distributions of the droplet mass caught at each sampling point are obtained by means of numerical simulation， so as to construct the sample space. Second， the POD method is utilized to find basis modes and the fitting coefficients needed to express and reconstruct the droplet mass caught. Finally， a surrogate model between each sampling point and the fitting coefficients is established by the Kriging model， the rapid prediction of the distribution of the droplet mass caught on the airfoil is developed. The validations of this estimation method show that the proposed method has an outstanding performance on accurate and rapid prediction of the distribution of droplet mass caught. Compared to the numerical simulation method， the time-consuming of the estimated method drops significantly， which will provide helpful support for unmanned aerial vehicle （UAV） anti-/de-icing design.