The frequency-based self-powered sensor has an important place in the era of the Internet of Things. In existing reports, the resonator is used either as an energy harvester or a sensor, yet it cannot undertake both demands simultaneously in a single structure. A parametric coupling device (abbreviated as PCD), capable of both temperature sensing and energy harvesting, is proposed with a common wings-inspired structure. Under one low frequency vibration excitation, the vibration energy can be potentially converted to electric energy, while the output frequency is observed to have a linear relationship with the temperature. The wings-inspired structure is comprised of three parts bonded in series, two side beams for collecting vibration energy one of them synchronously for temperature sensing, and a middle flexible beam for achieving parametric coupling. Once the heat is applied to the sensing one, the internal thermal stress is generated in the beam, while the reaction force acts on the anchor end of the beam to create compressive stress, for which the stiffness is decreased to cause frequency shifts and turn the working frequency domain. The parametric coupling is designed to improve the PCD performance, which is supported by theoretical analysis and experimental verification. With a manufactured MEMS PCD prototype, under a harmonic vibration excitation, when the temperature varies from 35.85 °C to 48.35 °C, both frequency shifts for sensing and working frequency domain for harvesting are 13.7 Hz which is 3.5 times larger than those of a non-parametric coupling device (abbreviated as NPCD). The sensing sensitivity of PCD is measured as 2.91 times higher than that of NPCD under a harmonic vibration excitation, while that of PCD is analyzed as 1.69 times higher than that of NPCD under a random one. Theoretical analysis shows the maximum temperature detection range is from 33.32 °C to 84.07 °C and can be further enlarged from 33.32 °C to 108.54 °C by adjusting the length of the sensing beam. The proposed PCD has the potential to be applied to temperature compensation systems or to maintain the operating environments of sensing nodes in self-powered wireless sensing networks (WSN).