Despite national curricula and instructional reforms calling for collaborative problem-solving skills (CPS), however, there is an absence of a theory-laden model showing how to effectively construct CPS for science learning. We therefore developed and validated a simulation-based CPS model that exploits its constructs, sequences, and causal relationships, and evaluating its effectiveness on students' problem-solving. Over the span of a two-week physics science course, 57 ninth-grade students were recruited from two intact middle school classes to engage in this online simulation-based collaborative problem-solving (CPS) program. This program consisted of nine electrochemistry problem-solving lessons spread across four class sessions, each lasting 45 min. Results indicated that the simulation-based CPS model was validated and proven to contribute to effective problem-solving by linking PS solution proposing, peer communication, implementing PS solutions with simulation, and providing evidence-based explanations. The simulation-based CPS model successfully improved the performance of both high- and low-achieving students. With the support and presence of high-achievers, low-achievers' collaboration attitude was boosted, which lead them to achieve similar learning success.