The aim of this study was to evaluate the feasibility and reproducibility of diffusion-weighted imaging in distinguishing between mucin-producing and serous pancreatic cysts. Forty-four pancreatic cysts (43 patients, 27 women; mean age, 57 years; 26 mucin-producing cysts, 18 serous cysts) that underwent histological examination or cyst analysis after diffusion-weighted magnetic resonance imaging were retrospectively reviewed. Three blinded readers independently evaluated signal intensity and apparent diffusion coefficient (ADC). Intraobserver and interobserver agreements were calculated. Fisher exact test and Welch t test were used to compare signal intensity and ADC values, respectively, with pathological results. Receiver operating characteristic analysis was used to determine diagnostic accuracy of various thresholds for ADC. A P value less than 0.05 was considered statistically significant. Mean ADC values of the mucin-producing cysts were 3.26 × 10, 3.27 × 10, and 3.35 × 10 mm/s for the 3 readers, respectively. Mean ADC values of the serous cysts were 2.86 × 10, 2.85 × 10, and 2.85 × 10 mm/s for the 3 readers, respectively. Differences in ADC values between the 2 cyst groups were 12.4%, 12.9%, and 14.8% for the 3 readers, respectively (P < 0.001). Intraobserver and interobserver agreement was excellent. A threshold ADC of 3 × 10 mm/s resulted in correct identification of cysts in 77% to 81% of cases, with sensitivity and specificity ranging between 84% and 88% and 66% and 72%, respectively. Diffusion-weighted imaging may be a helpful tool in distinguishing between mucin-producing and serous pancreatic cysts.