The present study investigated to what extent measured dissolved metal concentrations, WHAM-predicted free metal ion activity and modulating water chemistry factors can predict Ni, Cu, Zn, Cd and Pb accumulation in various aquatic insects under natural field conditions. Total dissolved concentrations and accumulated metal levels in four taxa (Leuctra sp., Simuliidae, Rhithrogena sp. and Perlodidae) were determined and free metal ion activities were calculated in 36 headwater streams located in the north-west part of England. Observed invertebrate body burdens were strongly related to free metal ion activities and competition among cations for uptake in the biota. Taking into account competitive effects generally provided better fits than considering uptake as a function of total dissolved metal levels or the free ion alone. Due to the critical importance and large range in pH (4.09 to 8.33), the H+ ion activity was the most dominant factor influencing metal accumulation. Adding the influence of Na+ on Cu2+ accumulation improved the model goodness of fit for both Rhithrogena sp. and Perlodidae. Effects of hardness ions on metal accumulation were limited, indicating the minor influence of Ca2+ and Mg2+ on metal accumulation in soft-water streams (0.01 to 0.94mM Ca; 0.02 to 0.39mM Mg). DOC levels (ranging from 0.6 to 8.9mgL−1) significantly affected Cu body burdens, however not the accumulation of the other metals. Our results suggest that 1) uptake and accumulation of free metal ions are most dominantly influenced by competition of free H+ ions in low-hardness headwaters and 2) invertebrate body burdens in natural waters can be predicted based on the free metal ion activity using speciation modelling and effects of H+ competition. •Strong relations were observed between body burdens and free metal ion activities.•The effect of H+ ions on insect body burdens was most clearly revealed.•Effects of major hardness ions on metal accumulation were rather limited.•Insect body burdens in natural waters can be predicted using speciation modelling.