The environmental risk assessment of metals in the soil compartment is based mainly on tests performed in Organization for Economic Cooperation and Development (OECD) artificial soil, but ecologically, the use of natural soils would be more relevant. In this contribution, the reproduction and growth of three standard species (an earthworm, a collembolan, and a dicotyledonous plant, respectively) was evaluated in nine natural soils (covering a wide range of pH values, organic matter content, texture, and so on) and in OECD artificial soil. Afterward, the effects of the model chemical zinc nitrate were assessed in all soils that were identified as being suitable for these species. The test results indicate that the toxicity of zinc nitrate can be higher by a factor of approximately four compared to artificial soil for invertebrates (earthworms and collembolans), whereas plants are only slightly more sensitive in some natural soils than in artificial soil. When comparing the different endpoints, it could be confirmed that the median effective concentration (EC50) is the most robust compared to the highly uncertain 10% effective concentration. Decreasing toxicity of zinc nitrate to collembolans was significantly correlated with an increase in soil pH but not with cation exchange capacity (CEC) or organic carbon (OC) content. No significant correlation was found between the toxicity of zinc nitrate to earthworms or plants and soil pH, CEC, or OC content. Possible consequences of the results are discussed, such as the testing of natural soils in addition to the OECD artificial soil or the inclusion of an additional safety factor to use the EC50 in current risk assessment schemes focusing on no-observed-effect concentrations.