Objective To evaluate the reproducibility of normal liver ADC measurements by different respiratory motion compensation techniques. Methods A total of 31 young healthy volunteers who are 20 to 40 years old without any hepatic diseases were selected to research. Each volunteer underwent liver DWI twice in 24 hours with the same parameters and location. The imaging was performed with free-breath(FB), breathhold(BH), rspiratory-triggered(RT)and navigator-triggered(NT)techniques, and the ADC values of the left hepatic lobe and right hepatic lobe (upper, middle and lower) was acquired with two scans. Analysis the the ADC values of various anatomic locations of liver with two-way analysis of variance of randomized block design. Reproducibility of ADCs was assessed with the Bland-Altman method. Analysis of variance and paired-sample t test was used to assess ADCs from both right and left liver lobe among the four techniques. Result The ADC values acquired from the four techniques were significant differences (P<0.01),and the ADC values of the right lobe were less than the left lobe's(P<0.01). It showed a trend to decrease moving from superior to inferior levels in both left and right lobes, and the ADC values among The middle and lower were significant differences (P<0.01). The limit of agreement of ADC of twice imaging among the four techniques were as follow: the right lobe was less than the left lobes, and the breathhold was less the others. As the result, reproducibility in the right liver lobe was better to that in the left and the reproducibility with breathhold was better than the other respiratory motion compensation techniques. Conclusions Both anatomic location and DWI technique influence the liver ADC measurements and their reproducibility. The reproducibility of BH is the best.