Many studies have evaluated the relationships between alkaline phosphatase (ALP) levels and the prognosis for osteosarcoma. However, a consensus has yet to be reached. We completed a meta-analysis to assess the significance of ALP and prognosis for osteosarcoma. We retrieved eligible documents from the PubMed and Embase databases and extracted related data from those documents. The overall survival (OS), hazard ratio (HR) and event-free survival (EFS) HR were obtained after combination to evaluate the impacts of ALP levels on prognosis for osteosarcoma. After screening, a total of 12 documents published between 1999 and 2013 were included. The ALP levels on OS were evaluated in nine documents. The pooled HRs was 1.78 (95% CI: 1.52-2.07, p < .05). The ALP levels on EFS were determined in eight documents. The pooled HRs was 1.58 (95% CI: 1.37-1.82, p < .05). Begg's test (OS, p > .754; EFS, p > .386) and Egger's test (OS, p > .649; EFS, p > .274) showed that there was no significant publication bias during analytic process. In summary, our meta-analysis shows that a higher level of ALP can decrease the OS and EFS in patients with osteosarcoma and ALP is an important biological indicator for patients with osteosarcoma.