This paper presents a method of modeling and solving a very complicated real logistic problem in the management of transportation and sales. The problem to be addressed is a large-scale multicommodity, multi-source and multi-sink network flow optimization, of 12 types of coal from 29 mines, through over 200 railway stations along 5 railroad arteries, in Chongqing Coal Industry Company of China. A minimum-cost flow model is established for the network system, and several maximal-flow algorithms are implemented to produce an optimal scheme.