Intelligence decision is an important subject for robotic intelligence system. This paper presents an experimental research work about robotic intelligence decision, where a fast information reduction with feedback features is proposed to solve robot intelligent decision and judgement issue. We have investigated the properties satisfied by the robotic information processing and decision process: if an attribute is the only discerning of multiple information objects, this attribute can be taken as the first core attribute, and can be reused in an iterative information reduction process to get another core attribute; furthermore, a feedback method can be used to enhance this iterative processing capability in real-time robots cooperative systems. Out feedback control method is somewhat dynamical and iterative process, and the investigated properties are utilized to design the feedback control laws. We have set up the relative experimental platform, and the experimental results show that our fast robot feedback decision algorithm is efficient and effective, which can be used in multiple robots collaborative missions.