© Proceedings of the 38th International Conference on Ground Control in Mining, ICGCM 2019. All rights reserved. Due to recent coal bursts in Australian coalmines, research of coal bursts has attracted intensive interest from mining researchers and engineers worldwide. Previous research has shown that the sudden drop of fractal dimension may be a precursor of catastrophic failure of rock. This paper examines the fractal characteristics of temporal-spatial distribution of acoustic emission signals released during the compression loading process of coal. This research suggests the possibility of coal burst forecasting based on acoustic emission data. The concept and calculation equation of fractal dimension is introduced in this paper. Additionally, uniaxial compression experiments of Australian coal samples are conducted, and the acoustic emission waveform and 3D location are recorded simultaneously by eight sensors Express-8 AE monitoring equipment in the loading process. The experimental results indicate that the continued slowdown of fractal dimension in time sequence can be used as a precursory signal of coal sample brittle failure. The fractal dimension of the spatial distribution of acoustic emission events decreases with the evolution of coal fracturing.