In this paper conventional data collection exercise was carried out to investigate the causes of equipment failure in Olefin plant using a series of interviews about the plants life for the past six years. The two variables used are the maintenance hour (X) and the loss in production during maintenance (Y). Mathematical approach of regression analysis was used to analyze the failure. The developed model was simulated to determine the intercept and the slop, which is the change in production loss that accompanies a unit change in maintenance hour. The correlation coefficient of maintenance hour and production loss and the standard error of the slope were computed. The simulated model reveals that the intercept is zero, the slope is 24, and the correlation coefficient is 1 and 95% prediction interval for the slope. Human error, depreciation and lack of maintenance after startup were considered as the contributing factors to the failure of industrial equipment.