Abstract Aiming at forecasting techniques of ground pressure hazards in deep metal mines, a microseismic monitoring system was established at a deep stope in Hongtoushan copper mine in China. Based on theories of geophysics and rock mechanics, a forecasting approach was introduced using microseismic multi-parameters including cumulative apparent volume, energy index, spatial correlation length, fractal dimension and b value. The approach was used to predict large scale fractures and ground pressure hazards in rock masses. The results showed that microseismic activities were very weak before mining commenced. However, activities became significant when mining started. When mining finished, microseismic activities remained at a higher level for about one week then became weakened. Before large scale fractures occurred, all microseismic multi-parameters behaved differently. The apparent volume and spatial correlation length were continually increasing, while energy index, fractal dimension and b value were gradually dropping to their minimum values. Therefore, these parameters could be used as precursory indicators for great ground pressure hazards. Comprehensive analysis on microseismic multi-parameters changes could improve reliability of forecasting, which is of great significance to deep metal mines in managing ground pressure failure risks.