To improve the operation lifetime of devices and the anti-interference capability of complex electromagnetic environments in device-to-device (D2D) networks, an energy-efficient maximization resource allocation algorithm was proposed for simultaneous wireless information and power transfer (SWIPT)-based D2D networks. Considering the constraints of users' minimum rate, minimum harvested energy, power splitting, and channel allocation, a robust energy-efficient resource allocation model was formulated by jointly optimizing the transmit power, subchannel coefficients, and the power splitting factor. The original NP-hard problem was converted into a deterministic and convex problem by using Dinkelbach's method and the worst-case approach, where the closed-form solutions were obtained by using the matching theory and Lagrange dual theory. Simulation results demonstrated that the proposed algorithm exhibited better robustness and higher energy efficiency.