This study investigates the rate of returns to private investment in education in urban China, focusing on gender differences. It shows that in general females have higher rates of returns to schooling than males after taking account of sample selection bias and the endogeneity of schooling, despite the fact that females usually have less schooling and lower income. However, the advances of females become less prominent after controlling for occupational choices. Furthermore, the sub samples of rural-to-urban migrant workers and urban-resident workers display different patterns: for urban residents, females have slightly higher rates of returns to schooling, while migrant workers show an opposite hierarchy of gender differences in returns to schooling, in the sense that males have higher returns to schooling than females.