Purpose <p>Soil respiration (Rs) is a critical component of the terrestrial carbon cycle, yet how degradation severity and vegetation restoration jointly regulate Rs dynamics at different analytical scales remains poorly understood. This study aimed to investigate the interactive effects of soil degradation severity and different vegetation restoration types on Rs and to identify the underlying scale-dependent regulatory mechanisms in Northeast China’s degraded Mollisols.</p> Materials and Methods <p>We conducted a field experiment using a factorial design with three degradation levels (lightly, LD; moderately, MD; severely, SD) and five treatments: <i>Populus simonii</i> × <i>P. nigra</i> (<i>P. xiaohei</i>), <i>Larix gmelinii</i> (<i>L. gmelinii</i>), <i>Amorpha fruticosa</i> (<i>A. fruticosa</i>), <i>Bromus inermis</i> (<i>B. inermis</i>), and bare soil control (CK). Rs was measured biweekly during the growing season. Complementary structural equation modeling approaches (seasonal integration: <i>n</i> = 540; end-of-season mechanistic assessment: <i>n</i> = 45) and Random Forest analysis were employed to characterize hydrothermal and biochemical regulatory pathways. Concurrently, soil physicochemical properties, enzyme activities, and plant growth metrics were analyzed to determine their relationships with Rs dynamics.</p> Results <p>Degradation severity systematically suppressed cumulative soil respiration during the growing season active period, with bare soil emissions declining from 28.68&#xa0;g C m^-2 (LD) to 4.23&#xa0;g C m^-2 (SD), representing an 85% reduction in seasonal carbon efflux. Vegetation restoration demonstrated stage-dependent recovery potential: in LD soils, restoration effects were minimal with most species showing similar cumulative Rs compared to bare soil. However, in MD soils, <i>P. xiaohei</i> achieved the highest increase (+ 77% vs. CK), while the most pronounced restoration effects occurred in SD soils, where <i>B. inermis</i> produced 107% enhancement and <i>A. fruticosa</i> increased Rs by 85% compared to CK, though neither approach fully restored emissions to LD levels. At seasonal scales, a critical shift occurs in dominant environmental drivers: temperature sensitivity (<i>Q</i>₁₀ &gt; 1) governs Rs in LD and MD soils, whereas severe moisture limitation (<i>Q</i>₁₀ ≈ 1) becomes the primary constraint in SD soils. End-of-season assessment identified a complementary biochemical pathway where degradation imposes persistent limitations through cascading effects: soil degradation degree → nutrient depletion → enzyme inhibition → reduced Rs.</p> Conclusions <p>Successful restoration requires stage-specific strategies matching plant functional traits to degradation-imposed constraints. Fast-turnover species (<i>P. xiaohei</i>, <i>B. inermis</i>) should prioritize MD soils to accelerate nutrient cycling, while water-adaptive species (<i>A. fruticosa</i>, <i>L. gmelinii</i>) are essential for SD soils to overcome moisture bottlenecks. However, vegetation-based approaches alone cannot fully restore severely degraded systems, necessitating integrated management combining strategic species selection with physical amendments and targeted nutrient rehabilitation.</p>