Studies have indicated that 25-hydroxyvitamin D (25(OH)D) level in obese is lower than normal weight subjects; however, results of studies that investigated relationship between 25(OH)D and fat mass are inconsistent. In addition, several randomized clinical trials (RCTs) have studied the influence of cholecalciferol supplement on percentage fat mass (PFM) but their results are conflicting. The objectives were to investigate the association between vitamin D3 and PFM pooling together observational studies and RCTs. PubMed/MEDLINE, Cochrane, and Scopus were comprehensively searched from inception to September 2016. The Fisher's Z (SE) of correlation coefficient and mean (SD) of changes in PFM from baseline were used to perform meta-analysis in observational studies and RCTs, respectively. To determine potential source of heterogeneity, subgroup and meta-regression analyses were conducted. Pooling correlation coefficients showed an inverse association between PFM (Fisher's Z: - 0.24, 95% CI: - 0.30 to -0 .18) and FM (Fisher's Z: - 0.32, 95% CI: - 0.43 to - 0.22) and 25(OH)D. Subgroup analysis revealed continent but not gender influence on the effect size. Meta-regression analysis indicated that age, latitude, and longitude are not sources of heterogeneity. Combining trials showed vitamin D3 supplementation had a mild but insignificant effect on PFM (- 0.31%, 95% CI: - 1.07 to 0.44). Subgroup analyses indicated that type of cholecalciferol and treatment regimens explain source of heterogeneity. Age, baseline body mass index, dose of cholecalciferol, length of study, female (%), and baseline 25(OH)D are not source of heterogeneity. In conclusion, our results state that 25(OH)D level is inversely correlated with PFM but cholecalciferol supplementation had no effect on PFM.