Abstract Background Hepatectomy with vascular resection (VR) for perihilar cholangiocarcinoma (PHCC) is a challenging procedure. However, only a few reports on this procedure have been published and its clinical significance has not been fully evaluated. Methods Patients undergoing surgical resection for PHCC from 2002–2017 were studied. The surgical outcomes of VR and non-VR groups were compared. Results Some 238 patients were included. VR was performed in 85 patients. The resected vessels were hepatic artery alone (31 patients), portal vein alone (37 patients) or both (17 patients). The morbidity rates were almost the same in the VR (49.4 per cent) and non-VR (43.8 per cent) groups (P = 0.404). The mortality rates of VR (3.5 per cent) and non-VR (3.3 per cent) were also comparable (P > 0.999). The median survival time (MST) was 45 months in the non-VR group and 36 months in VR group (P = 0.124). Among patients in whom tumour involvement was suspected on preoperative imaging and whose carbohydrate antigen 19-9 (CA19-9) value was 37 U/ml or less, MST in the VR group was significantly longer than that in the non-VR group (50 versus 34 months, P = 0.017). In contrast, when the CA19-9 value was greater than 37 U/ml, MST of the VR and non-VR groups was comparable (28 versus 29 months, P = 0.520). Conclusion Hepatectomy with VR for PHCC can be performed in a highly specialized hepatobiliary centre with equivalent short- and long-term outcomes to hepatectomy without VR.