We developed models to predict the effect of water velocity on prey capture rates and on optimal foraging velocities of two sympatric juvenile salmonids, coho salmon and steelhead. Mean fish size was ~80 mm, the size of age I+ coho and steelhead during their second summer in Southeast Alaska streams, when size overlap suggests that competition might be strongest. We used experimentally determined prey capture probabilities to estimate the effect of water velocity on gross energy intake rates, and we modeled prey capture costs using experimental data for search and handling times and published models of swimming costs. We used the difference between gross energy intake and prey capture costs to predict velocities at which each species maximized net energy intake rate. Predicted prey capture rates for both species declined from ~75 to 3040 prey/h with a velocity increase from 0.30 to 0.60 m·s1. We found little difference between coho and steelhead in predicted optimum foraging velocities (0.29 m·s1 for coho and 0.30 m·s1 for steelhead). Although prey capture ability appears to be more important than are prey capture costs in determining optimum foraging velocities, capture costs may be important for models that predict fish growth. Because coho are assumed to pay a greater swimming cost due to a less hydrodynamic body form, we also modeled 10 and 25% increases in hydrodynamic drag to assess the effect of increased prey capture costs. This reduced optimum velocity by 0 and 0.01 ms1, respectively. Habitat segregation among equal-sized coho and steelhead does not appear to be related to the effects of water velocity on their respective foraging abilities