<p>Current science standards emphasize the ability to use simulations to solve problems, but many students struggle to handle scientific tasks in simulation contexts, and little is known about the cognitive mechanism underlying this ability, which makes it difficult to teach. Therefore, the aim of this study was to generate a cognitive model of scientific problem-solving while using simulations. For this purpose, the think-aloud method and screen capture recordings were used to gather information from 27 middle school students as they tackled seven simulation-based scientific tasks. An inductive-deductive approach to analysis led to a coding scheme and a model revealing that the cognitive process comprised 20 components and five stages—identifying and extracting, designing and investigating, analyzing and arguing, monitoring and evaluating, and summarizing and clarifying—demonstrating a non-linear iterative cyclic process. Limitations and implications for future research and practices are discussed.</p>