<p>Global environmental change and human activities have widespread and profound impacts on vegetation growth, yet their temporal dynamics and relative contribution remain unclear, especially in large-scale ecological restoration projects. We used kernelized Normalized Difference Vegetation Index (kNDVI) data to analyze the vegetation trend in “Shan-shui Initiative” projects (SSIs) during 2000–2022 and utilized a refined random forest (RF) residual analysis to quantify the contributions of global environmental change and human activities before and after project implementation. The results indicated that the kNDVI in SSIs experienced a significant fluctuating increase (<i>slope</i> = 0.0021&#xa0;year⁻¹, <i>p</i> &lt; 0.01) over the period. However, the kNDVI trend shifted after project implementation. The proportion of areas with significant increases diminished from 56.8% to 28.0%, while the stable trend expanded substantially from 42.1% to 69.1%. The partial correlation between kNDVI and the five environmental factors exhibited significant spatial heterogeneity, with specific factors experiencing a shift in correlation following project implementation. Attribution results further demonstrate that before implementation, vegetation was more vulnerable to global environmental change (89.4% positive effects) than to human activities (63.8%). Following the launch of the SSIs, human activities emerged as the leading positive driver, with its positive contribution area (55.6%) surpassing that of global environmental factors, which decreased significantly to 50.6%. These findings provide comprehensive insight into the dynamic effects of global environmental change and human activities on vegetation trends, thereby offering a scientific foundation for adaptive regional management strategies.</p>