Dreissena rostriformis bugensis and Dreissena polymorpha, collectively known as dreissenid mussels, were first documented in Lake Erie nearly 40 years ago and continue to strongly impact ecosystem structure and function. Recently, dreissenids have been recognized as the primary regulator of phosphorus dynamics in the Great Lakes, potentially reducing external phosphorus load management strategies. To evaluate the extent that dreissenids may modulate phosphorus dynamics and assess the effectiveness of external load management strategies, ecosystem models need to adequately represent spatial patterns of dreissenid biomass densities. Assessment of dreissenid populations in the Great Lakes are complicated by their vast size, as well as the heterogeneity and stochastic nature of critical physical variables that shape habitat suitability (e.g., depth, hypoxia, substrate type, and benthic shear stress). We used observations from a 2019 benthic survey and 10-fold cross validation to assess spatial prediction methods to create a spatially explicit data layer representing dreissenid biomass in Lake Erie, including inverse distance weighting, nearest neighbour, depth binned average, random forest, generalized linear, and general additive models. A random forest spatial prediction approach produced the best predictive surface for dreissenid biomass in Lake Erie. The estimated spatial pattern of dreissenid biomass may be used in nutrient modeling exercises across Lake Erie to improve estimates of dreissenid impacts on biogeochemistry and lower food web productivity. This work also provides insight for potential sampling design improvements for benthic surveys in future years that could help refine the predictive capacity of the developed spatial prediction framework.