Background Spontaneous tumor rupture is a rare and life-threatening complication of hepatocellular carcinoma (HCC). The best treatment strategy remains unclear. Methods The clinical data of 137 patients with spontaneously ruptured HCC from 2010 to 2015 were reviewed retrospectively. We investigated the outcome and prognostic factors of various treatment strategies. Results Of the 137 patients, 53, 45, 3, and 36 patients underwent transcatheter arterial chemoembolization (TACE) alone, liver resection (LR) (LR alone or TACE + LR), surgical hemostasis, and conservative therapy. The patients undergoing LR had longest overall survival (OS). In the TACE alone group, independent factors affecting 30-day mortality were MELD score >= 12, AFP >= 1000 ng/ml, and largest tumor size >= 10 cm. AFP >= 1000 ng/ml, largest tumor size >= 10 cm, and no tumor capsule were significantly associated with poorer OS. In the LR group, largest tumor size >= 10 cm and no tumor capsule were the only independent prognostic factors for poorer OS and recurrence-free survival (RFS). Hypovolemic shock was an independent prognostic factor for poorer OS. The differences in OS between the TACE + LR group and LR alone group were not significant (P = 0.955). However, the RFS is significantly better in the LR alone group than those in the TACE + LR group (P = 0.031). Conclusion For resectable tumor, LR is the treatment of choice for patients with spontaneous ruptured HCC and preserved liver function. The delay in LR due to preoperative TACE may account for its worse RFS compared with LR alone. In patients with an unresectable tumor, TACE therapy alone improved survival over conservative therapy.