Background The optimal sampling techniques for endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) remain unclear and have not been standardized. The aim of this study was to compare the wet-suction and dry-suction techniques for sampling solid lesions in the pancreas, mediastinum, and abdomen. Methods This was a multicenter, crossover, randomized controlled trial with randomized order of sampling techniques. The 296 consecutive patients underwent EUS-FNA with 22G needles and were randomized in a ratio of 1: 1 into two separate groups that received the dry-suction and wet-suction techniques in a different order. The primary outcome was to compare the histological diagnostic accuracy of dry suction and wet suction for malignancy. The secondary outcomes were to compare the cytological diagnostic accuracy and specimen quality. Results Among the 269 patients with pancreatic (n = 161) and non-pancreatic (n = 108) lesions analyzed, the wet-suction technique had a significantly better histological diagnostic accuracy (84.9% [95% confidence interval (CI) 79.9%-89.0%] vs. 73.2% [95%CI 67.1%-78.7%]; P = 0.001), higher specimen adequacy (94.8% vs. 78.8%; P < 0.001), and less blood contamination (P < 0.001) than the dry-suction technique. In addition, sampling non-pancreatic lesions with two passes of wet suction provided a histological diagnostic accuracy of 91.6%. Conclusions The wet-suction technique in EUS-FNA generates better histological diagnostic accuracy and specimen quality than the dry-suction technique. Furthermore, sampling non-pancreatic lesions with two passes of EUSFNA with wet suction may provide a definitive histological diagnosis when rapid on-site evaluation is not routinely available.