The number of numerical models and their complexity are increasing rapidly. This not only accompanies with a bigger impact in various fields but also with a growing opacity. In the case of a deterministic model with hundreds of inputs, it can become difficult, while crucial, to identify which mechanisms are the most influential on the results. Screening sensitivity analysis methods, such as the Morris method, were developed to establish the ranking of inputs' influence on a specified output. The advantages of these methods lie in their easy adaptation to different model formulations or input sets and in providing accurate rankings at relatively low computational costs. In this work, a sensitivity analysis based on the Morris method is applied to a cold oxygen plasma model. The robustness of the method and some of its enhancements are tested, and the rankings obtained are compared with other existing ones for similar conditions. Some significant results are nonintuitive, which highlights the interest of using well-designed sensitivity analysis to study, control, and optimize complex systems.