Serum creatinine (Cre) is used as a surrogate marker of muscle mass. We investigated the impact of the Cre-to-bodyweight (BW) ratio (Cre/BW) on incident non-alcoholic fatty liver disease (NAFLD). Fatty liver disease was diagnosed by abdominal ultrasonography. In this historical cohort study of 13 728 participants (6397 men and 7331 women), we divided the participants into two groups by sex and into quartiles according to Cre (mg/dL)/BW (kg; Q1-4). We carried out Cox proportional hazard models, adjusting for age, alanine aminotransferase, fasting plasma glucose, systolic blood pressure, alcohol consumption, smoking status, and exercise. During the 5.1-year follow up for men and 6.0-year follow up for women, 2497 participants (1696 men, 801 women) developed NAFLD. The 4000-days cumulative incidence rates of FLD for men and women were 29.6% and 16.6% in Q1, 28.2% and 10.6% in Q2, 25.5% and 8.8% in Q3, and 22.7% and 7.7% in Q4, respectively. The hazard ratios of incident NAFLD in Q1 (Cre/BW [×100]: men <1.28; women <1.17) were 1.89 (95% confidence interval 1.64-2.17, P < 0.001) in men and 2.96 (2.42-3.62, P < 0.001) in women, compared with Q4 (Cre/BW [×100]: men ≥1.61; women ≥1.51). A low Cre/BW is associated with an increased risk of NAFLD. Screening Cre/BW can be used to identify individuals who are at high risk of NAFLD. © 2019 The Japan Society of Hepatology.