Background: Our previous study clarified that morbidity was a negative prognostic factor and sarcopenia defined by of the handgrip strength was a risk factor for the morbidity in gastric cancer surgery. Sarcopenia was reportedly a negative prognostic factor in colorectal cancer, hepatocellular carcinoma and malignant melanoma. This study aimed to evaluate impact of preoperative sarcopenia on recurrence-free survival (RFS) in gastric cancer surgery. Methods: Between May 2011 and June 2013, 256 consecutive primary gastric cancer patients who underwent curative surgery were retrospectively examined. Patients who received neoadjuvant chemotherapy or were diagnosed with pathological stage IV were excluded. Preoperative skeletal muscle mass was evaluated by bioelectrical impedance analysis and was expressed as skeletal muscle index or SMI (muscle mass/height²) by adjusting absolute muscle mass with height. Preoperative muscle function was measured by hand grip strength (HGS). Each cutoff value was determined as the gender-specific lowest 20% of the distribution of each measurement. Univariate and multivariate analyses were preformed to identify risk factors for RFS using a Cox proportional hazards model. Results: Median age (range) was 66 years (37-85 years). Male to female ratio was 168:88. Median follow-up period was 33.4 months. Pathological stage was I in 160, II in 48 and III in 48 patients. Univariate analysis showed that age, adjuvant chemotherapy, pT, pN, histological type, tumor size, total gastrectomy, low SMI and low HGS were significant risk factors for RFS. Multi-variate Cox’s proportional hazard analyses demonstrated that pT (HR 2.76, p = 0.0001), pN (HR 1.375, p = 0.037), histological type (HR 3.46, p = 0.014), low SMI (HR2.17, p = 0.036) were the significant risk factors for RFS. The three-year RFS was 89.1% in the patients with high SMI and 73.2% in those with low SMI (p = 0.007). Conclusions: Low SMI was an independent risk factor for RFS in Stage I-III gastric cancer. Low HGS, a risk factor for morbidity shown in our previous study, was not a risk independent factor for RFS. Preoperative sarcopenia as the short- and long-term outcomes has a value to be tested in the future prospective studies in gastric cancer surgery.