Plant microbiomes have important roles in plant health and productivity. However, despite flowers being directly linked to reproductive outcomes, little is known about the microbiomes of flowers and their potential interaction with pathogen infection. Here, we investigated the temporal spatial dynamics of the apple stigma microbiome when challenged with a phytopathogen Erwinia amylovora, the causal agent of fire blight disease. We profiled the microbiome from the stigmas of individual flowers, greatly increasing the resolution at which we can characterize shifts in the composition of the microbiome. Individual flowers harbored unique microbiomes at the operational taxonomic unit level. However, taxonomic analysis of community succession showed a population gradually dominated by bacteria within the families Enterobacteriaceae and Pseudomonadaceae. Flowers inoculated with E. amylovora established large populations of the phytopathogen, with pathogen-specific gene counts of >3.0 × 107 in 90% of the flowers. Yet, only 42% of inoculated flowers later developed fire blight symptoms. This reveals that pathogen abundance on the stigma is not sufficient to predict disease outcome. Our data demonstrate that apple flowers represent an excellent model in which to characterize how plant microbiomes establish, develop, and correlate with biological processes such as disease progression in an experimentally tractable plant organ.