Maternal infection is thought to increase the risk of non-affective psychosis including schizophrenia. However, observational studies have produced conflicting results and little is known about the importance of timing of infection in mediating subsequent risk. In this study, we carried out a meta-analysis of observational studies to investigate the risk of maternal infection and subsequent risk of non-affective psychosis. Using seven cohort studies, we found that maternal infection during gestation increased the risk of non-affective psychosis [relative risk (RR): 1.28 (95% CI:1.05-1.57, p = 0.02, I2 = 36%)]. A subgroup analysis identified that there was greater risk for schizophrenia alone [RR: 1.65 (95% CI:1.23-2.22, p = 0.0008, I2 = 0%)]. In addition, infection during the second trimester resulted in increased risk [RR: 1.63 (95% CI:1.07-2.48, p = 0.02, I2 = 7%)], whilst risk during the first and third trimesters did not meet statistical significance. This study highlights maternal infection in gestation as an important environmental risk factor for non-affective psychosis and our findings carry important implications for future disease prevention strategies.