Objective: To estimate the cumulative incidence of self-reported influenza vaccination ("vaccination coverage") and investigate predictors in HIV-infected women. Methods: In an ongoing cohort study of HIV-infected women in five US cities, data from two influenza seasons (2006-2007 n= 1209 and 2007-2008 n= 1161) were used to estimate crude and adjusted prevalence ratios (aPR) and 95% confidence intervals ([,]) from Poisson regression with robust variance models using generalized estimating equations (GEE). Results: In our study, 55% and 57% of HIV-infected women reported vaccination during the 2006-2007 and 2007-2008 seasons, respectively. Using data from both seasons, older age, non-smoking status, CD4 T-lymphocyte (CD4) count ≥200 cells/mm3, and reporting at least one recent healthcare visit was associated with increased vaccination coverage. In the 2007-2008 season, a belief in the protection of the vaccine (aPR=1.38 [1.18, 1.61]) and influenza vaccination in the previous season (aPR=1.66 [1.44, 1.91]) most strongly predicted vaccination status. Conclusion: Interventions to reach unvaccinated HIV-infected women should focus on changing beliefs about the effectiveness of influenza vaccination and target younger women, current smokers, those without recent healthcare visits, or a CD4 count <200 cells/mm3.