Based on the obversation data and detailed metadata archive recorded by 129 meteorological stations in the northeast of China, the homogeneity test and corrections are carried out to the monthly maximum and minimum temperatures by using the penalized maximum T test (PMT), and the corrections have been compared to the previous study (CHHT). The results show that the discontinuities of monthly temperature are prevalent which are mainly caused by data joint from different resoures and station relocation. There are 74 and 94 breakpoints detected respectively from monthly maximum and minimum temperatures. The spatial consistency of the monthly maximum and minimum temperatures has been improved and the spatial distribution of long-term trends is more reasonable after data homogenization. The trends of annual maximum and minimum temperatures have respectively increased from 0.1℃·(10 a)~(-1) and 0.22℃·(10 a)~(-1) to 0.17℃·(10 a)~(-1) and 0.33℃·(10 a)~(-1). Compared to the CHHT, temporal correlation coefficient between PMT and CHHT is above 0.9, reflecting the consistent climate warming. The spatial distribution of long-term trends is slightly improved and the trends of annual maximum and minimum tempe-rature calculated by PMT are 0.05℃·(10 a)~(-1) and 0.03℃·(10 a)~(-1) warmer than those by CHHT.