More and more information are needed in social life and commercial production, causing significant pressure on the sampling and too much time spent on signal sampling. Compressed sensing is one emerging hotspot in signal processing which employs a special sampling method to capture and represent compressible signals at a rate significantly below the Nyquist rate. In this paper, a Takagi-Sugeno-Kang (TSK) Model based on compressed-sensing sampling theorem is proposed for grinding power. It is further tested by using the actual production data, and the algorithm performance in grinding power model is also analyzed. The experiments show the validity and effectiveness of the proposed modeling method and its bright application foreground in other fields with similar features, such as power, metallurgy and so on. More and more information are needed in social life and commercial production, causing significant pressure on the sampling and too much time time on signal sampling. Compressed sensing is one emerging hotspot in signal processing which employs a special sampling method to capture and represent compressible signals at a rate significantly below the Nyquist rate. In this paper, a Takagi-Sugeno-Kang (TSK) Model based on compressed-sensing sampling theorem is proposed for grinding power. It is further tested by using the actual production data, and the algorithm performance in grinding power model is also analyzed. The experiments show the validity and effectiveness of the proposed modeling method and its bright application foreground in other fields with similar features, such as power, metallurgy and so on.