Earthquakes are totally unpremeditated events and cannot be fully anticipated in advance. However, by analyzing historical knowledge and formulating mathematical models will offer researchers with an additional legitimate estimate of future losses. Earthquakes are the most deadly disaster in the world. The purpose of present study is to estimate the earthquake casualties and resulted damage costs for possible future earthquakes in Pakistan. On the basis of examination new mathematical models have been developed for estimation of earthquake casualties and related damage costs through multiple linear regression using matrices with correlation coefficient α = 0.01. This study considers twenty-two most damaging earthquake that hit different regions of Pakistan from 1909 to 2017. The resulting models were multiple linear regression models explaining earthquake casualties and a total cost of damage through five independent variables magnitude, intensity, and depth of focus, location of epicenter and interval of earthquake
