The fault diagnosis of power equipment can be looked as pattern classification problem. Each fault of power equipment is corresponding to a set of traits. In this paper

according to fault status and correlated measured parameters of power equipment fault cases

a fault diagnosis decision-making table （i.e. "knowledge base"） is structured. By way of the decision-making table

the diagnosis rules is extract directly from fault patterns using rough set method. The extracted rules can be used to offer the effective fault diagnosis for power equipment. This paper brings forward a new approach to process diagnosis and give the more satisfied diagnosis service for power equipment in remote diagnosis by information network when some diagnosis data information offered by user is not complete.