Defective functional-group-endowed polymer semiconductors, which have unique photoelectric properties and rapid carrier separation properties, are an emerging type of high-performance photocatalyst for various energy and environmental applications. However, traditional oxidation etching chemical methods struggle to introduce defects or produce special functional group structures gently and controllably, which limits the implementation and application of the defective functional group modification strategy. Here, with the surface carboxyl modification of graphitic carbon nitride (g-C3N4) photocatalyst as an example, we show for the first time the feasibility and precise modification potential of the non-thermal plasma method. In this method, the microwave plasma technique is employed to generate highly active plasma in a combined H2+CO2 gas environment. The plasma treatment allows for scalable production of high-quality defective carboxyl group-endowed g-C3N4 nanosheets with mesopores. The rapid H2+CO2 plasma immersion treatment can precisely tune the electronic and band structures of g-C3N4 nanosheets within 10 min. This conjoint approach also promotes charge-carrier separation and accelerates the photocatalyst-catalyzed H2 evolution rate from 1.68 mmol h−1g−1 (raw g-C3N4) to 8.53 mmol h−1g−1 (H2+CO2-pCN) under Xenon lamp irradiation. The apparent quantum yield (AQY) of the H2+CO2-pCN with the presence of 5 wt.% Pt cocatalyst is 4.14% at 450 nm. Combined with density functional theory calculations, we illustrate that the synergistic N vacancy generation and carboxyl species grafting modifies raw g-C3N4 materials by introducing ideal defective carboxyl groups into the framework of heptazine ring g-C3N4, leading to significantly optimized electronic structure and active sites for efficient photocatalytic H2 evolution. The 5.08-times enhancement in the photocatalytic H2 evolution over the as-developed catalysts reveal the potential and maneuverability of the non-thermal plasma method in positioning carboxyl defects and mesoporous morphology. This work presents new understanding about the defect engineering mechanism in g-C3N4 semiconductors, and thus paves the way for rational design of effective polymeric photocatalysts through advanced defective functional group engineering techniques evolving CO2 as the industrial carrier gas.