Facing the faults of some prevailing views in the research of resource science of today, authors put forward the new points that the resource science is a problem-oriented discipline. As a problem-oriented discipline, resoure science is greatly different from the conventional sciences in the evolving background, in the object and main contents of the guiding theories, in the research approaches and in the discipline skeleton. The skeleton of resource science should be made up of three parts: analysis, explanation and management (design), in terms of three steps: understanding,explaining and solving resource problems.