Objective: Climate change is fueling the rapid range expansion of invasive species in freshwater ecosystems. This has led to mounting calls from natural resource managers for more robust predictions of invasive species distributions to anticipate threats to species of concern and implement proactive conservation and restoration actions. Here, we applied recent advances in fish sampling and statistical modeling in river networks to estimate the current and future watershed-scale spatial distribution of nonnative Smallmouth Bass Micropterus dolomieu. Methods: We integrated a spatial stream network (SSN) model of stream temperature, landscape environmental covariates, and Smallmouth Bass occurrence data based on environmental DNA (eDNA) detections to develop an SSN species distribution model (SDM) representing current Smallmouth Bass distributions in the Chehalis River, Washington State, a large coastal river basin of ongoing watershed-scale restoration. The SDM was informed by spatially intensive eDNA sampling from 135 locations in the main stem and major tributaries. We then applied downscaled climate change projections to the SSN SDM to predict Smallmouth Bass range expansion in the basin by late century. Result: We identified high levels of spatial autocorrelation at hydrological distances of ≤10 km in our eDNA data set, underscoring the importance of applying an SSN modeling framework. Stream temperature was identified as the most important environmental covariate explaining variability in Smallmouth Bass occurrence. Model predictions estimated that current suitable summer habitat for Smallmouth Bass habitat spans 681 km and is projected to nearly double by late century (1333 km) under a moderate climate change scenario. Current and future suitable habitat for Smallmouth Bass is prevalent in important tributaries for spring Chinook Salmon Oncorhynchus tshawytscha, a species of major conservation concern in the Chehalis River and more broadly along the Pacific coast. In both the main stem and tributaries, the SSN SDM predictions of the upstream leading edges of Smallmouth Bass closely align with (within 4.8 km) edges identified by spatially intensive eDNA sampling. Conclusion: Our study highlights the value of integrating SSN models with rapidly growing eDNA data sets for accurate and precise riverine fish distribution estimation. Our application provides crucial insights for anticipating the impacts of shifting invasive species on Pacific salmon Oncorhynchus spp. in a warming world.