Because of the complexity and fuzziness of the real world, it’s hard to build a dense knowledge system and reason in it with traditional methods. But man can deal with such tasks freely. Inspired by cognition and linguistics, a solution is advanced for reasoning dense knowledge in this paper. Objects and concepts are organized in the form of concept graph. Soaking the nodes in the graph until the result is represented in the graph the final graph can be the explanation of the scenario. With the naïve algorithm, monotonic scenario reasoning problem can be solved in dense knowledge environment.