The longwall mining is considered to be the best coal mining practices due to vast recovery of coal over other forms of underground as well as opencast mining methods. But the actual scenario is quite opposite in India where productivity form of the longwall mines lags far behind than the desired level. Irregular caving and sudden rock bursts, which are very hazardous for mine workers and equipment. Usually these are major problem faced by bulk of Indian longwall faces and which are due to the presence of thick sandstone beds as overlying strata. Thus to keep an eye on the rock bursts, it is very necessary to monitor the stressed zones in the hanging overlying strata above and behind the panel. Earlier Correlation Integral ‘C’ and Correlation Fractal Dimension ‘Dc’ has been very helpful in monitoring the stressed zones for several great earthquakes in past. Following the same way, in the present study we have used the mine-induced microseismic data obtained from the retreating longwall panel using various monitoring instruments to calculate the Correlation Fractal Dimension ‘Dc’ for monitoring the stress levels and fractures in the overlying strata and also for spatio-temporal forecasting of roof-falls. The variation of blast charge size with Fractal Dimension is also studied. The use of Fractal Dimension has been very effective in obtaining the precursory signatures for roof-fall, thus ensuring safety in the mines. •Identification of stressed zones using Fractal signatures.•Prediction of roof-fall both within and behind the panel.•Spatio-temporal forecasting of roof-fall using Fractal Dimension.•Estimation of blast charge size using Fractal Dimension.•Fracture monitoring in the overlying strata.