Three statistical models that predict microbial interactions within the vaginal environment are presented. A large data set was assembled from in vivo studies describing the healthy vaginal environment, and the data set was analyzed to determine whether statistical models which would accurately predict the interactions of the microflora in this environment could be formulated. During assembly of the data set, two new variables were defined and were added to the data set, that is, cycle (sequence of menstrual cycle) and flow stage (subdivision of cycle determined by day of menstrual cycle). Concentrations of total aerobic (includes facultative) bacteria, total anaerobic bacteria, and a Corynebacterium sp. were identified by correlation analysis as variables with significant predictors. By using a regression method with a backward elimination procedure, significant predictors of these outcome variables were identified as the concentrations of Lactobacillus spp., anaerobic Streptococcus spp., and Staphylococcus spp., respectively. For all three outcome variables, pH and flow stage were also identified as significant independent variables. Because some of the data in the data set are repeated measurements for a subject, a mixed-effect model that accounts for the random effects of repeated-measurement data fit best the data set for predicting interactions between various members of the vaginal microflora. The predictive accuracies of the three models were tested by a comparison of model-predicted outcome-variable values with actual mean in vivo outcome-variable values. From these results, we concluded that it is possible to accurately predict vaginal microflora interactions by using a mixed-effect modeling system. The application of this type of modeling strategy and its future use are discussed.