To make replacement decision, a decision maker must consider both the technological improvement and the heavy interactions among equipment in a manufacturing system. Thus, he (or she) must include all possible future equipment and all possible equipment combinations in generating replacement alternatives, which makes the number of alternatives very large. To present all these alternatives effectively, we suggest using a configuration network. And we proposes a pseudo-dynamic programming (PDP) approach and a multi-attribute decision criterion named weighted satisfaction level (WSL) to select the best alternative that maximizes the long term profit of a company.