Lyme disease risk is related to the abundance of infected nymphal ticks, which in turn depends on the abundance and reservoir competence of wild hosts. Reservoir competence of a host (i.e., probability that an infected host will infect a feeding vector) often declines over time after inoculation, and small mammalian reservoirs typically undergo rapid population growth during the period when vector ticks feed. These processes can affect disease risk in the context of site-specific tick abundance and host community composition. We modeled the effects of reservoir decay and host demographic turnover on Lyme disease risk using a simple yearly difference equation model and a more realistic simulation incorporating seasonal dynamics of ticks and hosts. Both reservoir decay and demographic turnover caused (1) specific infectivity (proportion infected 3 reservoir competence) of host populations to vary with host community composition, (2) tick infection prevalence and the specific infectivity of reservoirs to be highly sensitive to the abundance of questing nymphs, and (3) specific infectivity and the infection prevalence of ticks to decrease at high host densities. Reservoir competence decay had similar effects in both model formulations, but host turnover had less effect than reservoir decay in the seasonal model. In general, exponential reservoir decay and abrupt loss of reservoir competence had similar effects, although exponential decay caused greater sensitivity to tick density and host community composition. Reservoir decay may explain the observed variability in published field measurements of reservoir competence of a host species. Our results illuminate mechanisms by which host diversity can dilute the impact of a highly competent reservoir and suggest that management to reduce nymphal tick abundance may reap an added benefit by reducing nymphal infection prevalence.