Toxicity data for microorganism in soil or in soil less cultures have been described with ion competition models, however these models disregard electrostatic and osmotic effects which are known to affect ion sorption and toxicity. Using European soils with diverse characteristics, the factors that influence the toxicity of soil Cu or Ni to potential nitrification rate (PNR) and glucose-induced respiration (GIR) were evaluated based on the electrical potential (ψ0) and ion activities ({M2+}0) at the outer surfaces of bacterial cell membranes (CMs). The zeta potentials (ζ) of bacterial (Escherichia coli) protoplasts, as affected by the ionic composition of the solution, were measured and used to estimate the parameters of a Gouy–Chapman–Stern (GCS) model which was then used to compute ψ0 values. The ψ0 values varied widely with soil type and increased markedly (became less negative) as metal salts were added. Computed ψ0 was then used to predict the surface ion activities from the soil solution composition. The toxicity data (both PNR and GIR) were statistically related to (i) surface activities of free metal ions ({M2+}0), (ii) the ameliorative effect of surface H+ activity ({H+}0), (iii) the ψ0-influenced electrical driving force for cation uptake across CMs, and (iv) osmotic effects. This electrostatic model predicted the observed GIR and PNR with Radj2>0.816 for observed vs. predicted PNR and Radj2>0.861 for observed vs. predicted GIR. These predictions were generally better than those by previous models. The suggestion that metal toxicity in spiked soils is partly related to a spike-induced osmotic increase is corroborated by fitting the model to spiked soils that were or were not leached and aged to reduce the osmotic increase. The predicted soil EC50 values (in mg metal/kg soil) were within a factor of 2.5 for up to nineteen European soils with a wide range of properties.