Objective To assess the efficacy of metformin in megestrol acetate (MA)‐based fertility‐sparing treatment for patients with atypical endometrial hyperplasia (AEH) and endometrioid endometrial cancer (EEC). Design A randomised, single‐centre, open‐label, controlled trial conducted between October 2013 and December 2017. Setting Shanghai OBGYN Hospital of Fudan University, China. Population A total of 150 patients (18–45 years old) with primary AEH or well‐differentiated EEC were randomised into an MA group (n = 74) and an MA plus metformin group (n = 76). Methods Patients with AEH or EEC were firstly stratified, then randomised to receive MA (160 mg orally, daily) or MA (160 mg orally, daily) plus metformin (500 mg orally, three times a day). Main outcomes and measures The primary efficacy parameter was the cumulate complete response (CR) rate within 16 weeks of treatment (16w‐CR rate); the secondary efficacy parameters were 30w‐CR rate and adverse events. Results The 16w‐CR rate was higher in the metformin plus MA group than in the MA‐only group (34.3 versus 20.7%, odds ratio OR 2.0, 95% confidence interval CI 0.89–4.51, P = 0.09) but the difference was more significant in 102 AEH patients (39.6 versus 20.4%, OR 2.56, 95% CI 1.06–6.21, P = 0.04). This effect of metformin was also significant in non‐obese (51.4 versus 24.3%, OR 3.28, 95% CI 1.22–8.84, P = 0.02) and insulin‐sensitive (54.8 versus 28.6%, OR 3.04, 95% CI 1.03–8.97, P = 0.04) subgroups of AEH women. No significant result was found in secondary endpoints. Conclusion As a fertility‐sparing treatment, metformin plus MA was associated with a higher early CR rate compared with MA alone in AEH patients. Tweetable For AEH patients, metformin plus MA might be a better fertility‐sparing treatment to achieve a higher early CR rate compared with MA alone. Tweetable For AEH patients, metformin plus MA might be a better fertility‐sparing treatment to achieve higher early CR rate compared with MA alone.