The relationship between hospital volume and patient outcome is globally known; thus, hospital volume is widely used as a quality indicator. In Japan, however, recent studies on this topic are scarce. The present study examined whether hospital surgery volume is associated with postoperative 5‐year survival among cancer patients. Using the Osaka Cancer Registry, we identified a sample of 86 145 patients who were diagnosed with cancer at any of five different sites (stomach, colorectum, lung, breast and uterus) and underwent surgeries between 2007 and 2011 in Osaka. We ranked hospitals by annual surgical volume, sorted patients in descending order by hospital volume, and assigned them into quartiles (high, medium, low and very low volume). We analyzed the association between hospital volume and 5‐year survival among 80 959 patients aged between 15 and 84 years using Cox proportional hazard models. Adjustments were made for characteristics of patients, type of surgery and adjuvant treatment received. The mortality hazard of patients treated at very low‐volume hospitals was 1.36‐1.82‐fold higher than that of patients treated at high‐volume hospitals. Absolute differences in adjusted survival rates between high‐volume and very low‐volume hospitals varied with the cancer site: 14.9 in stomach, 11.5 in colorectal, 10.8 in lung, 2.4 in breast and 3.3 in uterine cancers. Hospitals with lower surgery volumes showed higher mortality risks after cancer surgery than those with higher volumes. Monitoring site‐specific surgery volumes and referring patients from low‐volume to high‐volume hospitals may be beneficial for improving the long‐term survival of cancer patients.