Elevated serum phosphate is associated with cardiovascular and total mortality in patients with kidney diseases and healthy individuals. But whether serum phosphate is associated with stroke is controversial. We searched PubMed and Embase from January 1, 1970 to May 9, 2014 with keywords such as "serum phosphate", "serum phosphorus", and "stroke". Dose-response meta-analysis was conducted. A restricted cubic spline model was used, and then we estimated pooled RR using generalized least square regression taking into account the correlation among categories of each study. Five studies with a total of 32,608 patients were included. We identified a linear relationship between serum phosphate and risk of stroke (P for non-linearity = 0.5258). The RR of phosphate (1 mg/dL) for stroke was 1.00 (95 % CI 0.97-1.05), similar results were observed in subgroup analysis. A linear relationship between serum phosphate and risk of stroke is identified. There is no association between serum level of phosphate and stroke.