搜集整理2003~2022年中国30个省份的面板数据，采用双重机器学习的因果推断模型和面板门槛模型。研究发现：首先，环境规制和绿色创新能够双向驱动绿色经济增长，并释放积极有效的交互溢出效应。其次，环境规制持续强化过程中，绿色创新对绿色经济增长的溢出弹性系数具有先升后降的∩形变化特征；反向来看，绿色创新能力持续提升过程中，环境规制对绿色经济增长的溢出影响从不显著变为正向显著，并且在高阶区间呈现边际递增的动态演化规律。最后，环境规制与绿色创新双向驱动绿色经济增长的交互溢出效应在空间层面有所差异，环境规制的最优溢出门槛阈值呈现东高西低的分布特征；绿色创新的最优溢出门槛阈值则呈现东低中高的分布特征。

The paper collects panel data of 30 provinces in China from 2003 to 2022, using DDML models and panel threshold models. It is found that, firstly, both environmental regulation and green innovation can drive green economic growth in bidirectional effectively, and there is positive and effective interactive spillover effect between them. Secondly, during the process of increasing environmental regulation intensity, the elasticity coefficient of the impact of green innovation on green economic growth exhibits a ∩ shaped evolutionary characteristic from rising to falling. Reversely, during the process of enhancing green innovation capabilities, the spillover effect of environmental regulations on green economic growth has changed from insignificant to positively significant, and shows a dynamic spillover law of marginal increase in higher intervals. Finally, the interactive spillover effect of environmental regulation and green innovation on green economic growth exhibits spatial heterogeneity, the optimal spillover threshold for environmental regulation intensity shows the distribution characteristics of higher in the east regions, and lower in the west regions. While the optimal spillover threshold for green innovation capability shows the distribution characteristics of higher in the middle regions, and lower in the east regions.