The aim of this paper was the evaluation of the ecotoxic effects of three sulphonylurea compounds, commonly used as herbicides against weeds in agriculture by mathematical models. The tests were performed both in laboratory and field conditions to study the changes in soil urease, considered a key enzyme for regulating soil nitrogen transformation and a sensitive indicator to herbicides. Mathematical models accurately reproduced the behavior of urease from chemically treated soil samples, based on real, observable processes. Besides, they simplify the view regarding the activity of the selected enzyme, as there are many factors and complicated biogeochemical processes, which might interfere. Overall, we conclude that, for the cambic chernozem model analyzed here, the normal (label-recommended) chlorsulfuron and amidosulfuron doses do not perturb soil urease activity and the former compound is more urease-friendly than either amidosulfuron or tifensulfuron. In the context of the long-term use of these herbicides, our research underlines the importance of mathematical models and the prefiguration of a map for the differentiation of field / laboratory experiments, for the most accurate highlighting of the biochemical imbalances caused by the chemical substances, the risk of overdose and the toxicity risk for soil and environment.