Frequent climate changes and intense human activities increase the risk of vegetation destruction. Spectral index-based change detection can identify vegetation destruction from multi-temporal images, providing valuable insights for vegetation management and post-disaster recovery efforts. However, we still face the challenge of the spectral diversity of vegetation destruction and the complexity of the background environment. Existing spectral indices (VIs) often struggle to accurately detect vegetation destruction in complex scenarios. These VIs focus on specific aspects of vegetation, such as leaf, canopy or water content, limiting their effectiveness in capturing vegetation dynamics. In addition, they are susceptible to background environment changes. To overcome these challenges, this study proposes a new metric called Slope Vegetation Index (SVI) using bands that are sensitive vegetation leaf, canopy, and water content (i.e., green, near-infrared (NIR), and short-wave infrared (SWIR) bands). The performance of SVI was verified by the dual time-phase difference method, and five widely used VIs were selected for detailed comparison. In addition, the performance of SVI was evaluated using PROSAIL simulation data, various vegetation change scenarios, and real vegetation destruction cases. Moreover, we assessed the applicability of SVI to other multispectral sensors. The results showed that compared with existing VIs, SVI exhibited the highest sensitivity to vegetation changes under different chlorophyll and water content conditions. In various vegetation change scenarios and vegetation destruction cases, SVI consistently had the best performance, with Producer’s Accuracy (PA), User’s Accuracy (UA), and F1 scores all exceeding 0.90. In complex scenarios, SVI could better highlight vegetation changes while suppressing background environment changes. Additionally, SVI performed well on other Landsat-8/9 images, with an F1 score exceeding 0.89. This study confirms that SVI is valid for vegetation destruction detection and has potential for large-scale and high-frequency vegetation monitoring.