Experiments are crucial for sustainability science because they allow researchers to produce evidence about the causes of sustainability problems and about the effectiveness of solutions. Many laboratory and field studies, community-based initiatives, and pilot projects have been defined as experiments in this field. Yet, in sustainability science, it is still unclear what distinguishes scientific experiments from conventional projects or initiatives as well as how different scientific experiments compare to one another. In this article, we define an experiment as a scientific practice that relies primarily on an intervention and that allows for the production of empirical evidence. We show that, in sustainability science, researchers can have different types of control over the intervention (from full to no control) and that evidence can be about different subjects (sustainability problems or sustainability solutions). Relying on this differentiation, we introduce a typology that organizes experiments in sustainability science according to type of control over interventions and subjects of experimentation. The typology provides a synthetic, comprehensive, and comparative overview of the variety of experimental approaches in sustainability science. By providing a definition and a typology for scientific experimentation in this field, the article contributes to the further development of evidence-based approaches in sustainability science.