Vast active faults and their seismic activity in Yunnan-Sichuan area had been developed since the Cenozoic. Based on the earthquake catalogue and existing data of active faults, we compared the actual seismic activity of major fault zones in the past ten years in Yunnan-Sichuan area with the forecast results made by GSHAP (Global Seismic Hazard Assessment Program), and then concluded the similarities and differences between them.  The comparison showed that seismic activities in Longmenshan fault zone were greatly underestimated in GSHAP, for most of the earthquakes along Longmenshan fault zone were aftershocks of the 2008 Wenchuan earthquake. Coulomb stress change caused by the mainshock of Wenchuan earthquake triggered the 2013 Lushan earthquake, and then reduced the seismic hazard of the southeastern segment of Xianshuihe fault. In the past decade, low seismic activities showed in outer arc belt of Yunnan-Sichuan area, mainly including Xianshuihe fault, Anninghe fault, Xiaojiang fault and other minor faults, whose seismic hazard were greatly overestimated in GSHAP. The similar situation appeared on Red River fault, Lancangjiang fault and Litang fault, part of the inner arc belt, which had the highest seismic hazard assessed by GSHAP. Otherwise, seismic activities of Yingjiang area in southwestern Yunnan started to enhance since 2008, while the seismic hazard in that area was neglected in GSHAP. Along Nujiang fault, only in the south segment and Baoshan area existed higher seismic activity, which was consistent with GSHAP. On the other side, eastern piedmont fault of the Haba-Yulong Snow Mountains, belonging to the middle part of the inner arc belt, showed higher seismic activities in recent years, which conformed to the assessment of GSHAP. From the Zhaotong area to the east of Xiaojiang fault showed high seismic activities, consistent with GSHAP as well. The above comparisons were based on the earthquake catalogue in the past ten years, while the time horizon of GSHAP would be fifty years in future, and cycle of one devastating earthquake could be millennium, therefore large difference doesn't mean complete distortion of the estimates of GSHAP. 