The goal of this study was to develop a Korean version of the Global Physical Activity Questionnaire (K-GPAQ) and to examine its reliability and validity. The English version of the GPAQ was translated to the Korean language (K-GPAQ) via forward–backward translation. Reliability of the K-GPAQ was evaluated using a one-week interval test–retest method with 115 individuals. Criterion-related validity of the K-GPAQ was examined with 199 participants using accelerometers. Cohen’s kappa and Spearman’s correlation coefficients were used to measure test–retest reliability and validity, respectively. A Bland–Altman analysis was used to assess agreement between physical activity (PA) levels measured via K-GPAQ and the accelerometer. Coefficients for the reliability of the K-GPAQ showed moderate agreement for recreational PA and slight agreement for work-related PA (Cohen’s kappa: 0.60–0.67 for recreational PA and 0.30–0.38 for work-related PA and Spearman’s rho: 0.27–0.47 for work-related PA and 0.53–0.70 for recreational PA). Criterion validity of the total amount of PA, as measured by the K-GPAQ and the accelerometer, showed a weak but significant correlation (r = 0.34, p < 0.01). The K-GPAQ is a reliable and valid questionnaire to measure PA although K-GPAQ overestimated PA levels.