The purpose of this study was to evaluate the efficacy of taxane regimens in patients with metastatic angiosarcoma. Forty-one patients with metastatic angiosarcoma treated at the National Cancer Center Hospital between January 1982 and January 2009 were retrospectively classified into 3 groups according to the treatment type: (i) taxane (n=11), (ii) non-taxane (n=14), and (iii) best supportive care (BSC; n=16) groups. The taxane group received paclitaxel (n=6), docetaxel (n=4), or albumin-bound paclitaxel (n=1), and the non-taxane group received mainly doxorubicin-containing regimens (n=12). The differences in progression-free survival (PFS) among the 3 groups were statistically significant (P<0.001). After adjusting for prognostic factors, the taxane group had significantly longer PFS than the non-taxane (hazard ratio=0.282; 95% confidence interval=0.086-0.923; P=0.036) and BSC (hazard ratio=0.015; 95% confidence interval = 0.003-0.083; P<0.001) groups. Overall survival was also significantly longer in the taxane group than in the other groups. A taxane regimen may be more effective than a non-taxane regimen for treating patients with metastatic angiosarcoma.