A contrastive study of the classical degree-day model and the enhanced temperature-index model, incorporating the hourly global radiation, were present on Koxkar Glacier, a mountain glacier on the south slopes of the Tianshan Mountains, in the summer, 2008. The two models were run with three different spatial resolutions of 200 m, 100 m and 50 m contour interval, to test their applicability to glacial ablation with low data requirement. Hourly solar radiation was estimated by using parameterizing transfer theory, combind with FY 2C and NCEP/NCAR data. The performance of simulated and measured incoming shortwave radiation showed good consistent relationship, with a determination coefficient R2=0.74. Ablation computations of three sub-periods, using the classical degree-day model, indicated evident temporal variation of degree day factors over the whole summer. With spatial resolution increased, the both models can give good results, especially when the spatial resolutions was 50 m contour interval. It found that the enhanced temperature-index model can get better result than the simple degree-day model with spatial resolutions of 200 m and 100 m contour interval. However, there was no evident improvement by using the enhanced temperature-index model when the spatial resolution was 50 m contour interval due to the uncertainty of simulated incoming shortwave radiation.