Pandemic influenza is an international public health concern. In light of the persistent threat of H5N1 avian influenza and the recent pandemic of A/H1N1swine influenza outbreak, public health agencies around the globe are continuously revising their preparedness plans. The A/H1N1 pandemic of 2009 demonstrated that influenza activity and severity might vary considerably among age groups and locations, and the distribution of an effective influenza vaccine may be significantly delayed and staggered. Thus, pandemic influenza vaccine distribution policies should be tailored to the demographic and spatial structures of communities. Here, we introduce a bi-criteria decision-making framework for vaccine distribution policies that is based on a geospatial and demographically-structured model of pandemic influenza transmission within and between counties of Arizona in the Unites States. Based on data from the 2009-2010 H1N1 pandemic, the policy predicted to reduce overall attack rate most effectively is prioritizing counties expected to experience the latest epidemic waves (a policy that may be politically untenable). However, when we consider reductions in both the attack rate and the waiting period for those seeking vaccines, the widely adopted pro rata policy (distributing according to population size) is also predicted to be an effective strategy.