We work in complex industrial system simulation on geographical, decisional and informational distribution terms. We postulate that existing methodologies and simulation tools seldom take account of these three distribution aspects. Our goal is to conceive a modeling and simulation solution that is able to utilize these aspects to resolve problems. This paper is divided into 4 parts. The first is a brief introduction. The second is a short resume of the state of the art in the distributed simulation and the multi- gent system fields. There, we present the bases used for our reasoning. The second part describes the industrial system distribution modeling problems. In this section, we also develop the methological multi-agent approach. Thus the use of distributed simulation, economical and multi-agent works enable us to consider a model on which is based our approach. Throughout in this section, we present an iterative construction approach of our methodology and simulation environment. Finally, we conclude.