Promoting sexual abstinence among never-married youth is an important component of HIV/AIDS prevention campaigns for youth in countries with generalized epidemics. The objective of this study was to examine trends in primary and secondary abstinence among never-married youth age 15-24 in Kenya over a ten-year period and to explore the role of HIV prevention knowledge, schooling, and contextual factors in affecting abstinence behavior. Data were from Kenya Demographic and Health Surveys conducted in 1993, 1998, and 2003. Primary abstinence was defined as never-married youth who never had sex. Secondary abstinence was defined as never-married youth who ever had sex but not in the past year. Logistic regression models were used to estimate the effects of prevention knowledge, schooling, and contextual factors in affecting abstinence behavior, after accounting for sampling weights and clustering in the survey design. Both primary and secondary abstinence levels have risen in the past 10 years in Kenya. The abstinence levels were higher among female youth than among male youth. Multivariate analyses show that knowledge that abstinence can prevent HIV infection was positively associated with the likelihood of practicing abstinence (both primary and secondary). However, knowledge that condom use can prevent HIV infection was associated with lower abstinence practice. In-school youth were 4-5 times more likely to abstain from sex than those working (aOR=4.12; p=0.000 for female youth and aOR=4.83; p=0.000 for male youth). Not-in-school female youth were about as likely to abstain as working female youth, but not-in-school male youth were about 2 times more likely to abstain than working male youth. Muslim youth were much more likely to abstain than other youth. Female youth with weekly exposure to television and those with a secondary or higher education were significantly more likely to have abstained, whereas male youth were significantly less likely. Effects of these factors on secondary abstinence practice were generally weaker, but sexually experienced in-school male youth were significantly more likely to have abstained in the past year than sexually experienced working male youth. Effects of the contextual variables on the likelihood of abstinence were generally small and insignificant, except for primary abstinence among female youth. Conclusions: Increasing knowledge that abstinence can prevent HIV infection and keeping youth in schools can help promote abstinence behavior. Abstinence programs need to be gender sensitive and culturally appropriate.