There is limited evidence on the hypothesis that maternal occupational exposure near conception increases the risk of cancer in offspring. This study is to investigate whether women employed in an electronics factory increases childhood cancer among first live born singletons. We linked the databases of Birth Registration and Labor Insurance, and National Cancer Registry, which identified 40,647 female workers ever employed in this factory who gave 40,647 first live born singletons, and 47 of them developed cancers during 1979-2001. Mothers employed in this factory during their periconceptional periods (3 months before and after conception) were considered as exposed and compared with those not employed during the same periods. Poisson regression model was constructed to adjust for potential confounding by maternal age, education, sex, and year of birth. Based on 11 exposed cases, the rate ratio of all malignant neoplasms was increased to 2.26 [95% confidence interval (CI), 1.12-4.54] among children whose mothers worked in this factory during periconceptional periods. The RRs were associated with 6 years or less (RR=3.05; 95% CI, 1.20-7.74) and 7-9 years (RR=2.49; 95% CI, 1.26-4.94) of education compared with 10 years or more. An increased association was also found between childhood leukemia and exposed pregnancies (RR=3.83; 95% CI, 1.17-12.55). Our study suggests that maternal occupation with potential exposure to organic solvents during periconception might increase risks of childhood cancers, especially for leukemia.