To comprehensively exploit the rich spectral information and spatial texture features in multispectral remote sensing images， a change detection method based on Gabor-CVA is proposed to extract the change range and identify the change type. For two remote sensing images with different time phases in the same area， on the one hand， Principal Component Analysis （PCA） is applied to reduce dimensionality. Gabor kernel function is used to extract multi-scale and multi-directional spatial texture features and calculate the angle between texture vectors. On the other hand， Change Vector Analysis （CVA） is used to calculate spectral difference vector and spectral change intensity. Subsequently， on the fused feature image incorporating texture vector angles and spectral change magnitudes， thresholds are set to determine the distribution of change extents. Samples of land feature change types be selected， and the Random Forest （RF） method be employed to filter out several important change features from texture differences， spectral differences， and remote sensing index differences. Multiscale segmentation be performed on the corresponding change feature images. A Random Forest model be trained based on the segmented images and samples， and the change types of objects within the change extents be identified， thereby generating a map illustrating the distribution of change types. The experiment demonstrates that the Gabor-CVA method achieves an overall accuracy of 88.77% in detecting change ranges. Compared to the CVA method， under the condition of a 3.06% reduction in omission error， the true positive rate has increased by 21.56%. In comparison to the change detection method after RF classification， there is a reduction of 5.48% in omission error and an increase of 16.25% in true positive rate. In the classification of change types， the Kappa coefficient for the CVA method is 0.38. While the Kappa coefficient for change detection after RF classification is 0.50. Whereas the Gabor-CVA method significantly elevates the Kappa coefficient to 0.75.