Clinically, whether and how to make a surgical interventional decision for the patients with spinal metastases is still controversial. Life expectancy is a significant determinant in the selection of surgical procedure for spinal metastases. This study was to evaluate Tomita and Tokuhashi scoring systems in selecting surgical procedure and predicting prognosis of extradural spinal metastases. A total of 169 patients with spinal metastases, treated in the Spine Unit of Aarhus University Hospital, Denmark, from Jan. 2001 to Apr. 2004, were enrolled. The life expectancy was scored according to both Tomita system and Tokuhashi system before operation, the spinal metastases were classified according to Tomita system, and the patients underwent surgery accordingly. Follow-up was done 6, 12, and 24 months after operation. The precise of Tomita system and Tokuhashi system in estimating "death within 3 months", "death within 6 months", and "death within 12 months" was compared using Receiver Operating Characteristic curves (ROC curves). The mean survival time of the patients was calculated by Kaplan-Meier method. ROC curves of "death within 3 months", "death within 6 months", and "death within 12 months" showed no significant difference between Tomita score and Tokuhashi score in each group (P = 0.16, P = 0.47, and P = 0.38, respectively). Kaplan-Meier survival curves showed that Tomita system overestimated the prognosis in scores from 4 to 7, and Tokuhashi system underestimated the prognosis in scores from 0 to 8. Both Tomita and Tokuhashi scoring systems could be used to predict prognosis of spinal metastases after operation. Tokuhashi scoring system can predict early death more accurately, which can be used to avoid major operation for these patients.