Abstract: Coal mass is often affected by consecutive loading-unloading cycles in coal mining, so the study of energy conversion and failure mechanism of coal mass under repeated loads has a guiding significance for understanding the mine dynamic disasters. By using MTS815. 03 servo experimental system, the uniaxial loading cycle tests were performed for the mechanism of energy accumulation, dissipation, and release in the failure process of coal sample under different loading rates, and the fragmentation distribution was also analyzed based on the theoretical analysis of energy and fractal. It provides a basis for the study of burst response and failure mechanism of coal and rock under cyclic loading. The experimental results show that the phasic characteristics of energy conversion are obvious, which can be divided into three stages:initial energy accumulation stage, accelerated energy accumulation stage and fast energy dissipation stage. The proportion of energy dissipation of coal samples encounters the course from high to low again to high. The elastic energy has an opposite trend of change, the proportion of elastic energy drops or the proportion dissipated energy increases at the later stage of loading, the coal samples enter the accelerated failure stage. The energy dissipation and the energy release are closely related with the loading rate. With the increase of loading rate, the proportion of elastic energy increases before the peak stress, sample is damaged severely after the peak stress and its macro failure mode transits from shear tension and splitting failure to ejection failure. Because more energy is stored in the coal sample body in the form of elastic energy before the coal sample is destroyed, and more energy is released after the rock is destroyed, which makes the coal sample destroyed more severely. The phasic characteristics of fractal features of coal samples are obvious under cycling loading and unloading and the self-similarity characteristics are obvious within the range of threshold, the fractal dimension is between 2-3, which increases linearly with the loading rate. The fractal dimension of fragments increases with the increase of loading rate, and the higher the broken degree of coal samples, the smaller the proportion of large fragments. In this case, the higher of the broken level of more fragments and the smaller the mass of single fragments is, the greater the risk of dynamic disasters of coal samples.