With the proposal of a new type of power system dominated by new energy sources, the integration of energy technology and digital technology has further deepened, and the power grid is accelerating its evolution into an energy Internet. Building an energy Internet and its digital twin system imposes higher requirements on the power Internet of Things (IoT) simulation verification and operation platform and greatly expands the connotation and extension of the power IoT. This paper mainly focuses on how to build a power IoT simulation verification and operation platform for the digital twin system and analyzes the key technologies for the power IoT simulation verification and operation platform oriented toward the digital twin system. Subsequently, this paper proposes an evaluation strategy for a power IoT simulation verification and operation platform based on deep belief networks (DBN). Firstly, this paper introduces an adaptive step size adjustment mechanism and a fly population diversity adjustment mechanism using the classical fruit fly optimization algorithm (FOA) and proposes an improved FOA. Then, aiming at the optimization of the DBN structure parameters, the improved FOA algorithm is applied to optimize the initial structure parameters of DBN, in order to enhance the network performance. Additionally, the improved fruit fly optimization (IOFA)-deep belief networks (IFOA-DBN) algorithm for evaluating the power IoT simulation verification and operation platform is constructed. Finally, this paper conducts multi-level experiments to verify the effectiveness and correctness of the proposed method.