Abstract: The wireless microseismic sensor and acoustic emission monitoring technology developed for rock mass collapse monitoring were used to simulate the gradual natural stress in the process of rock mass collapse propagation by means of graded creep loading, and the acoustic signal characteristics of the laboratory simulation test process of dolomite crack collapse were studied. The results show that: for microseismic signals, the combined denoising method combining complete ensemble empirical mode decomposition (CEEMDAN) and wavelet is effective and can significantly improve the signal-to-noise ratio; after creep loading of each stage, microseismic signals are abundant in the initial stage of each creep loading, the amplitude and number of microseismic signals in the last creep loading both increase significantly, and the dominant frequency of microseismic signals decreases significantly before failure. The k value of acoustic emission signal(AF/RA) is defined and used as the criterion of tensile fracture and shear fracture. The proportion of shear fractures increase remarkably in the last five seconds to failure, especially in the last second to failure. The results of rock mass deformation monitoring in the test are compared and analyzed. It shows that the response of deformation monitoring to the whole process of loading failure of small-scale rock is weak, and it is difficult to extract effective failure precursor signals. The research results can provide rital support for the development of monitoring technology for rock mass collapse and instability.