Anti-smoking campaigns are widely viewed as a success case in public health policy. However, smoking rates continue to vary widely across U.S. states and the success of anti-smoking campaigns is contingent upon states' adoption of anti-smoking policies. Though state anti-smoking policy is a product of a political process, studies of the effect of policies on smoking prevalence have largely ignored how politics shapes policy adoption, which, in turn, impact state health outcomes. Policies may also have different effects in different political contexts. This study tests how state politics affects smoking prevalence both through the policies that states adopt (with policies playing a mediating role on health outcomes) or as an effect modifier of behavior (tobacco control policies may work differently in states in which the public is more or less receptive to them). The study uses publicly available data to construct a time-series cross-section dataset of state smoking prevalence, state political context, cigarette excise taxes, indoor smoking policies, and demographic characteristics from 1995 to 2013. Political ideology is measured using a validated indicator of the ideology of state legislatures and of the citizens of a state. We assess the relationship between state political context and state smoking prevalence rates adjusting for demographic characteristics and accounting for the mediating/moderating role of state policies with time and state fixed effects. We find that more liberal state ideology predicts lower adult smoking rates, but that the relationship between state ideology and adult smoking prevalence is only partly explained by state anti-smoking policies.