The paper considers theoretical and practical aspects of constructing a software architecture of energy efficient management plans for MIMO process facilities on a set of functioning states. The classification of control systems takes into account changes in functioning states during operation. The authors describe the possibilities of identifying the current functioning state in a control time slot. They also structure the software of energy efficient management plans with allocating functional subsystems and software modules. The general architecture of a energy efficient management plan includes a knowledge and data management subsystem, an interface subsystem and six basic software modules. The knowledge and data management subsystem includes a knowledge base, an inference engine, a database and a database management system. The interface subsystem consists of an initial data input module, a cognitive graphics module and an integrated development environment. The basic software modules of the energy efficient management plan are the following: an identification module of a control object dynamics model, a module for analysis of optimal control tasks, a module for synthesis of optimal control actions, a simulation module, a module for identification of a current functioning state, and an experiment planning module. Each basic module has determined system classes on a set of functioning states, in the software architecture of which it can be included. The paper proposes a methodology for constructing a software architecture of different energy efficient management plans on a set of functioning states. It also considers the features of software implementation of energy efficient management plans based on various approaches (using applied and software tools).