The relationship between fertility and the HIV/AIDS epidemic is not well understood. Although existing studies elsewhere generally point to the epidemic resulting in fertility reduction, earlier evidence from the Kenya Demographic and Health Survey 2003 (Central Bureau of Statistics [CBS], Kenya Ministry of Health [MOH] & ORC Macro, 2004), hereafter referred to as KDHS, showed interesting patterns, with regions most adversely affected with the HIV/AIDS epidemic showing the clearest sign of a reversal trend in fertility decline. HIV/AIDS may influence fertility through one or more behavioral and/or biological proximate fertility determinants. In this study, we explore: (i) the regional variations in the link between HIV/AIDS and fertility; (ii) possible mechanisms through which HIV/AIDS may influence fertility; and (iii) the effect of individual and contextual community-level HIV/AIDS factors on fertility. The study is based on secondary analysis of the 2003 KDHS data, which provides a unique opportunity to explore the impact of the HIV/AIDS epidemic on the affected populations, being the fourth survey in the international DHS program to include HIV testing, and the first to anonymously link the HIV results with key behavioral, social, and demographic factors at individual and household level. Multilevel models are used to examine the effect of individual and contextual community-level HIV/AIDS factors on fertility. The modeling is carried out in stages, starting with the key variables relating to HIV/AIDS, before introducing various proximate fertility determinants in successive stages, to explore possible mechanisms through which HIV/AIDS may influence fertility. The study corroborates findings of earlier studies on the fertility inhibiting effect of HIV/AIDS among infected women. HIV/AIDS infected women have 40 percent lower odds of having had a recent birth than their uninfected counterparts of similar background characteristics and child mortality experience. After taking into account proximate determinants of fertility relating to sexual exposure, breastfeeding duration, and fetal loss, the odds for HIV/AIDS infected women are 33 percent lower, suggesting that the effect of HIV/AIDS on fertility is partly through these proximate determinants. However, there is no evidence of a significant association between community level HIV/AIDS prevalence and fertility when the background socio-cultural and demographic factors are controlled for. The results suggest that although recent trends in sexual exposure factors (e.g. rising age at first sex and age at first marriage and a decline in the proportion of women in union) might be expected to sustain a declining trend in fertility, trends in some of the proximate determinants, including reduced duration of breastfeeding and increased child mortality coupled with reduced desire to stop childbearing may have contributed to the stalled fertility decline in Kenya. Whilst HIV/AIDS may have influenced the recent changes in sexual exposure factors, it is also likely that it has contributed to increasing infant and child mortality and reduced duration of breastfeeding, which are partly responsible for the stall in fertility decline. The regional patterns show that the most notable increase in fertility and the greatest decline in contraceptive prevalence were observed in Nyanza province, the region with the highest HIV/AIDS prevalence. The regional patterns of the other proximate determinants with respect to sexual exposure factors, infant/child mortality and duration of breastfeeding all show unfavorable patterns for the region. For instance, Nyanza has consistently recorded the lowest age at first sex, the lowest age at first marriage, and the highest infant and child mortality in Kenya across years. The recent trends in Nyanza have not been encouraging either: it witnessed the least overall rise in age at first sex and first marriage during the 1993-2003 period; and recorded among the greatest declines in the duration of breastfeeding. These patterns are likely to have contributed to the observed reversal of fertility decline in the region.