In this study, a methodology is proposed towards development of all uncertainty model that includes randomness in the occurrence of days-lost accidents in a coal mine. The accident/injury data consists of 1390 days-lost accident cases recorded at GLI-Tuncbilek underground lignite mine from January 1994 to December 2002. In the first step of proposed methodology, the frequency and the severity of the accidents have been modeled statistically by fitting appropriate distributions. The test done by BestFit software yields a chi-square value of 21.53 (p = 0.089) with 14 degrees of freedom and estimates the parameter of lambda for Poisson distribution as 12.87 accidents/month. For the severity component, a lognormal distribution is fitted to days-lost data and chi-square goodness-of-fit test calculates a value of 40.44 (p = 0.097) with 30 degrees of freedom. The parameters of lognormal distribution are estimated as a mean of 14.3 days and standard deviation of 23.1 days, respectively. Then, two distributions are basically combined by Monte Carlo simulation in order to construct relative risk levels in yearly base referring to the final cumulative distribution. Finally, a simple forecasting modeling is carried out in order to quantitatively predict the expected risk levels by using decomposition technique in time series analysis. Stochastic model estimates that although, there Would be substantial reduction in the expected number of accidents in the near future, the higher level of risks still should be a concern for the mine management. (C) 2007 Elsevier Ltd. All rights reserved.