In a conceptual design phase, engineers need to estimate, as accurately as possible, the cost of system developed from each concept (cost of concept). If engineers overestimate cost, engineers may be forced to unnecessarily cancel system development. On the other hand, if engineers underestimate cost, the final system may not be as profitable as initially expected. When engineers make a critical decision to select a concept, the dilemma is that there is typically no or limited information to accurately estimate cost of concept. This paper proposes a belief-based approach for estimating a cost of system in a conceptual design phase. In the proposed approach, engineers benchmark existing systems in the market that are similar to the new system. For these benchmarked systems, engineers estimate costs by subtracting profit margins from observed prices. Engineers develop a distribution of cost for the new system from these estimated costs.