Objective To evaluate rapid on-site evaluation (ROSE) in endoscopic ultrasound-guided fine needle aspiration (EUS‑FNA) for pancreatic solid lesions, and to compare the difference in ROSE performance between cytopathologists and trained endoscopists. Methods A total of 168 consecutive patients with pancreatic solid lesions who underwent EUS‑FNA from January 2014 to December 2020 at Fuding Hospital, Fujian University of Traditional Chinese Medicine were recruited. The patients who did not receive ROSE from January 2014 to November 2017 were included in N‑ROSE group (n=67). Since December 2017, the patients who intended to receive EUS‑FNA were divided into E‑ROSE group (n=59, patients who received EUS‑FNA and ROSE by endoscopists trained with cytopathology) and C‑ROSE group (n=42,patients who received EUS‑FNA by untrained endoscopists and ROSE by cytopathologists) according to random number table. The number of punctures, sample adequacy, cytological diagnosis, final diagnosis and diagnostic efficiency (including the sensitivity, the specificity, the positive predictive value, the negative predictive value and the accuracy) in 3 groups were compared. Results (1) The puncture number in N‑ROSE group (4.22±0.76) was significantly more than E‑ROSE group (3.12±0.79, P<0.001) and C‑ROSE group (3.24±0.91, P<0.001). (2) The proportions of adequate samples in N‑ROSE group [82.09% (55/67)] was significantly lower than those of E‑ROSE group [96.61% (57/59), χ2=5.308,P=0.021] and C‑ROSE group [97.62% (41/42), χ2=4.541,P=0.033]. The proportion of negative cytological diagnosis in N‑ROSE group [40.30% (27/67)] was significantly higher than those of E‑ROSE group [20.34% (12/59), χ2=5.848, P=0.016] and C‑ROSE group [19.05% (8/42), χ2=5.348,P=0.021]. (3) The sensitivity of N‑ROSE group [74.07% (40/54)] was significantly lower than those of E‑ROSE group [94.00% (47/50), χ2=6.151, P=0.013] and C‑ROSE group [94.44% (34/36), χ2=4.817,P=0.028]. The accuracy in N‑ROSE group [79.10% (53/67)] was significantly lower than those of E‑ROSE group [94.92% (56/59), χ2=5.433,P=0.020] and C‑ROSE group [95.24% (40/42), χ2=4.155, P=0.042]. (4) There was no significant difference in any observational index between E‑ROSE group and C‑ROSE group (P>0.05). Conclusion ROSE in EUS‑FNA can improve sample adequacy, the diagnostic sensitivity and accuracy, and reduce the number of punctures. The sample adequacy and diagnostic efficiency of endoscopists trained with cytopathology are comparable to those of cytopathologists.