Abstract: Liquid loading is a serious problem for normal production of gas wells in Yakela-Dalaoba deep condensate gas reservoir. The applicability of common methods for predicting liquid loading still needs to be demonstrated. The applicability of critical rate method and kinetic energy factor method is evaluated and opimized. The fundamental principles of each method for predicting liquid loading and differences among them are introduced. Different dynamic distribution of critical rate and surface tension along wellbore are considered in order to improve original critical rate models. Through error analysis, the methods suitable for calculating temperature and pressure distribution in high and low gas?liquid ratio wellbores of Yakela-Dalaoba gas reservoir are optimized respectively, and such methods suitable for predicting liquid loading are given through field case study. Results show that modified pseudo single?phase coupling model is suitable for calculating temperature and pressure distribution in high gas?liquid ratio wellbore, while Hagedorn?Brown method and Hasan method are suitable for it in low gas?liquid ratio wellbore of Yakela-Dalaoba gas reservoir. The improved critical rate models increase precision of liquid loading prediction compared to original models. The improved Li Min model and kinetic energy factor method that sets threshold value as 6 are of the highest precision, and both of them are suitable for predicting liquid loading of gas wells in Yakela-Dalaoba deep condensate gas reservoir