The satellite-based precipitation is a key data source for investigating regional precipitation,especially over ungauged area,which can improve the knowledge of understanding pattern of precipitation. However,due to the limited spatial coverage of calibration data,for the widely used Tropical Rainfall Measuring Mission (TRMM) Multi-satellite Precipitation Analysis (TMPA) 3B42 data,the new Climatological Calibration Algorithm (CCA) fails to effectively implement in the mid-high latitude regions,which leads to much more uncertainties and errors exist in TMPA. Two representative regions with similar climate and topography are selected, which located inside and outside the coverage of TRMM satellites (38°N,38°S),respectively. The similarity and difference in data accuracy between two regions are analyzed by four statistical indexes to reveal the effect of CCA on TMPA,based on observation data from weather stations during 1998-2012. The results indicate that there is a general overestimation of precipitation at daily,monthly and annual scale in both regions. Particularly, precipitation event ranging from 0 and 10 mm is severely overestimated. Though CCA can not get rid of this overestimation, the degree of overestimation is alleviated to some degree. It is clear that the degree of overestimation is higher in Liaoning with 8. 87% than that in Shandong with 2. 79%. In terms of variation of precipitation, the monthly and annual precipitation can be regenerated by TMPA. TMPA follow the similar trends to the observed. It is noted that,an unsatisfactory accuracy is shown for two regions with large fluctuations in annual variation during winter. But the application of CCA in TMPA in Shandong,makes an improvement in correlation between two precipitation datasets and weaken the degree of overestimation of precipitation by TMPA.