Objective: Both the Department of Health (DOH) and Ministry of Interior (MOI) of Taiwan have established a computerized death registry based on the death certificates of the dead in Taiwan during the past two decades. The purpose of this study is to verify the accuracy of the dataset of death registration for the DOH with the dataset from the MOL .Because the data of the DOH during 1980-84 lacked the identification number (ID no.), our second objective is to establish guidelines for replacing the missing ID no.. Method: Firstly, decedents in the two datasets linked by same ID no. and demography including dates of birth, death, and places of living between 1985 and 1997 were identified and codes of education and village taken from the MOI were filed in those of the DOH. Decedents that could only be linked by ID no. or demography between 1985 and 1997, were refiled the demography or ID no. after making sure of the correction in the MOI. Secondly, due to the lack of information of ID no. in the DOH for decedents between 1980 and 1984, we linked the two decedents with social demography including dates of birth and death, area codes of detailed address, and then imputed ID no. into dataset of the DOH based on dataset of the MOI. Result: There were 1,259,600 decedents identified from 1985-97 with completely the same ID no. and demographic data. After verification of the data, we refiled the missing demographic data for 3,168 subjects and the missing ID no. for 33,036 subjects into the dataset of the DOH. According to the study results, there were 0.0512% of total subjects with complete overlapping demographic data. Thus, after linking the two datasets with the same complete demographic data, we were able to refile 88.83% of the missing Ids for all the death certificate data of the DOH during 1980-84. There were 46,305 cases in the dataset of the DOH and 63,157 cases in that of the MOI during 1980-97, that were unlinkable. This was probably because they were lost track of or there was a delay in registration. Conclusion: This study corrected many errors, filled in ID no. for 1980-84, and imputed information of education and village into the death registration dataset of the DOH. This can be useful for future studies. Beside, the unlinkable cases in the datasets of the DOH and MOI were identified and a separate dataset files for future usage was built up.