Blast fishing is a widespread destructive practice that has long-term negative impacts on coral reef ecosystems. To better understand the long-term impact of blast fishing on these ecosystems within Southern Tanzania, in situ surveys were undertaken within a two-year timeframe across six reef sites in Mikindani Bay. These sites varied in the time since blast fishing was last undertaken (<5 years, 5–10 years, or >10 years). While there were limited differences in major benthic categories (soft substrate, algae, and hard coral), multivariate benthic community mix varied significantly with the time since last blast, which explained approximately 25% of the observed variation. This compositional shift was characterized by prominent transitions among major benthic genera including branching Acropora, Euphyllia, massive Porites, and Xenia. This finding suggests that community composition varies across areas with different histories of blast fishing. Coral communities at sites with a more recent history of blasting (<5 years) were largely characterized by a high relative proportion of massive Porites, while sites with intermediate times since blasting (5–10 years) showed a distinct dominance of Euphyllia. In contrast, sites with the longest time elapsed since blast fishing ceased (>10 years) were distinguished by two main taxa; Xenia and branching Acropora. These results suggest that the long-term composition of coral assemblages is highly sensitive to historical blast fishing disturbances. The effective prevention of blast fishing in Mikindani Bay, Tanzania and coral reefs more broadly is therefore critical to conserve the remaining cover, diversity, structure and biological functions of these marine habitats.