To evaluate the abilities of population-based birth defects surveillance system (PBBDSS) and hospital-based birth defects surveillance system(HBBDSS).We used capture-mark-recapture method(CMR) to analyze the data of the two kinds of birth defects surveillance systems from 1 Oct. 2006 to 31 Dec. 2006 in a county of Hunan province. Data from PBBDSS were defined as the first source and data from HBBDSS were defined as the second source.49 and 28 birth defect cases were found from PBBDSS and from HBBDSS respectively. Among these cases, 20 were marked. With the method of CMR, the estimated birth defects cases were 68 (95% CI: 56-70). The coincident rates of PBBDSS and HBBDSS were 72.1% and 41.2%, while the total coincident rates was 83.8% and the coincident rates from different sources was 57.1%. The unreported rates of PBBDSS and HBBDSS were 27.9% and 58.8%.Not only the HBBDSS but also the PBBDSS appeared to have had high unreported rates, suggesting that we could use CMR to adjust the rate of birth defects from the birth defects surveillance data.