This paper investigates the effect of education on the timing of marriage among Kenyan women and the relative effects of education across generations of women. Data used is drawn from the 1998 Kenya Demographic and Health Survey. The Cox proportional hazard model and linear regression analysis are used to analyse the data. The results show that education has a statistically significant and strong positive effect on a woman’s age at first marriage; the effect remaining robust in the presence of a number of controls. The highly educated women are more likely to delay marriage. Significant variations in the effect of education across the generations of women are apparent. The effect is greater for the younger women, indicating increased postponement of marriage. Premarital sexual activity, premarital childbearing, region of residence, religion and year of birth are also significantly associated with age at first marriage