Chain coal pillars are parts of the structure of longwall mining system that play a significant role in the stability of the entries. With mechanization and developments in the various aspects of the method, higher efficiency in optimization of the design of chain coal pillars seems appropriate. In this paper, the three main methods of chain pillar design, namely the empirical, analytical and numerical methods are compared. Real data from the Tabas-Iran coal mines have been used in order to make the comparison process reliable. It is concluded that the most apparent advantage of the empirical method is the reliability of the results while the use of numerical methods enjoys the advantage of flexibility. On the other hand, the analytical methods are complex unless simplifying assumptions are made that can substantially decrease the accuracy of result which is thought to be the main advantage of the design method. A new method is therefore introduced here that combines all of the three presently used methods and by doing so, the new method has the advantages of the three while minimizing complexities and inaccuracies associated with the use of the individual methods.