High-performance resonant mass sensors are increasingly demanded for the trace detection of volatile organic compounds (VOCs) and hazardous gases in environmental monitoring. A critical challenge in current gas sensors is the inherent trade-off between increasing the specific surface area for gas molecule adsorption and maintaining the structural stiffness required for high-frequency resonance. This paper proposes a novel high-sensitivity micro-resonant gas sensor based on a porous functionally graded material (FGM) sandwich microplate. To accurately calibrate the sensor's dynamic response at the micro-scale, a refined mathematical model is established using the Modified Couple Stress Theory (MCST) with a position-dependent material length scale parameter. Furthermore, a multi-objective topology optimization framework based on the Non-dominated Sorting Genetic Algorithm II (NSGA-II) is developed. Rather than using pre-defined porous patterns, the algorithm autonomously aligns the mass-removal regions with the physical neutral surface of the microplate. By maximizing mass reduction in this low-stress region while preserving flexural rigidity, the optimized free-topology design achieves a 29.9% increase in the fundamental frequency compared to conventional O-type designs. This frequency boost directly translates to a significantly enhanced mass responsivity, providing a rigorous theoretical and design foundation for next-generation trace gas measurement devices.