When searching for nature rules we encounter a fundamental difficulty: for technical reasons we are not able to investigate the whole Nature - we have to explore its sections i.e. empirical systems (structures). However, even then it might happen that the system is too complex and its comprehensive, direct investigation is impossible. In this case we can study the system by means of its abstract model e.g. a mathematical model. Abstract models of empirical systems usually present simplified modelled systems, where simplification depends on different reasons, e.g. on the purpose of model construction or the state of our knowledge on a modelled system. As a rule we also exclude from the model those independent variables, whose influence on a dependent variable is insignificant. Simplification does not have to mean that the validity of the simplified heuristic model is lower than the validity of the same model before its simplification. Some remarks refer to certain methodological problems associated with models and modelling, as well as scientific research performed with those tools.