The network level access control policy is based on policy rule. The policy rule is a basic building of a policy based system. Each policy contains set of conditions and actions. Here conditions are evaluated to determine whether the actions are performed. The existing work is based on packet filtering scenario. Here every policy can be translated into canonical form. That uses the “First Matching Rule” resolution strategy. The access control matrix is proposed to translate the policy. The Generalized Aryabhata Reminder Theorem (GART) is used for to construct the access control matrix. In this access control matrix rows represent users and columns represent files. In which each user is associated with key and each digital file is associated with lock.