Abstract: Cognitive radio technology, which can improve the utilization rate of spectrum resources, has broad application prospect, and is likely to lead to another major revolution in the field of wireless communications. Spectrum sensing is the first step of cognitive radio technology, but its performance is far from the actual needs. Developed a truly practical application of high-performance spectrum sensing method has become an urgent problem to be solved. In recent years, a new theory closely related to wireless communications, random matrix theory (RMT), has been further applied to achieve a high performance in the field of spectrum sensing. New algorithms based on RMT and the latest research rusults, non-asymptotic random matrix theory (Non-Asymptotic RMT), are expected to be the true high-performance sensing technology to practical applications. Therefore, the research on spectrum sensing technology based on random matrix theory has very important theoretical and practical value. This paper summarizes the development process of RMT in areas of cognitive radio spectrum sensing; introduces its basic principle, summed up its main characteristics, and prospects for further research.