OBJECTIVE. Endoscopic ultrasonography (EUS)-guided fine-needle aspiration (EUS-FNA) may facilitate tissue sampling for histopathological diagnosis of subepithelial tumors (SETs) in the gastrointestinal (GI) tract. However, immunohistochemistry is not always feasible using EUS-FNA samples due to the low quality of specimens often obtained by aspiration. This study aimed to compare the use of 22-gauge (G) EUS-guided fine-needle biopsy (EUS-FNB) with 22G EUS-FNA for core sampling used for histopathological examination, including immunohistochemistry, in patients with GI SETs. METHODS. Twenty-eight patients with GI SETs ≥2 cm in size were prospectively enrolled at five university hospitals in Korea between January and June 2013. They were randomized to undergo either EUS-FNB or EUS-FNA. RESULTS. A total of 22 patients was finally analyzed in this study: 10 and 12 patients underwent EUS-FNA and EUS-FNB, respectively. Compared to the EUS-FNA group, the EUS-FNB group had a significantly lower median number of needle passes to obtain macroscopically optimal core samples (4 vs. 2, p = 0.025); higher yield rates of macroscopically and histologically optimal core samples with three needle passes (30% vs. 92%, p = 0.006; 20% vs. 75%, p = 0.010, respectively); and a higher diagnostic sufficiency rate (20% vs. 75%, p = 0.010). No technical difficulties were encountered in either group. CONCLUSIONS. This study shows that EUS-FNB has a better ability to obtain histological core samples and a higher diagnostic sufficiency rate than EUS-FNA and that EUS-FNB is a feasible, safe, and preferable modality for adequate core sampling for histopathological diagnosis of GI SETs.