The aim of this study was to evaluate current practice in gastric ulcer follow-up to establish diagnostic yield and predictors of malignancy. Repeat gastroscopy is routinely performed to confirm gastric ulcer healing and exclude malignancy. However, the incidence of malignancy at follow-up endoscopy is low, without consensus regarding case selection and timing. New gastric ulcers diagnosed on gastroscopy at 2 institutions in Australia were identified through keyword search of endoscopy reports over a 5-year period (2013 to 2017). Data collected included patient demographics, clinical presentation, and endoscopic and histologic findings from initial and subsequent gastroscopies. Of 795 patients, repeat gastroscopy was performed in 440 (55%). Malignancy was diagnosed in 52 (7%) with 83% identified at initial gastroscopy. Eight cancers were identified at repeat gastroscopy with malignancy yield of 2% (8/440). Three were diagnosed in patients with benign initial ulcer histology (3/286, 1%). One cancer was diagnosed during follow-up in a patient with benign histology but no repeat gastroscopy (1/286, 0.3%). Predictors of benign ulcers were absence of endoscopic suspicion [odds ratio (OR) 0.1 (0.03-0.13), P≤0.005], complete healing on repeat gastroscopy [OR 0.5 (0.34-0.70), P=0.036] and benign initial histology [OR 0.12 (0.43-0.90), P≤0.005]. Seven percent of new gastric ulcers were malignant with most identified with biopsy on initial gastroscopy. Malignancy yield from follow-up gastroscopy was 2%. Diagnostic yield of endoscopic follow-up may be low in ulcers with benign appearance and adequate histology. However, current practice of repeat gastroscopy is warranted in the absence of patient-based and lesion-based predictors of malignancy.