Hyperspectral remote sensing technology has been widely used in vegetation type mapping, and the integrated use of spectral and texture features of remote sensing data has become an effective way to improve the accuracy of vegetation classification, but the Hughes effect generated by the high-dimensional feature space composed of multi-scale texture features and hyperspectral features reduced the image classification accuracy, and the randomness of feature optimization algorithms such as Genetic Algorithm (GA) can also cause uncertainty in image classification accuracy. In order to solve those problems, a high-dimensional spectral-texture feature space was constructed by spectral features of ZY1 02D AHSI hyperspectral data and multi-scale texture features extracted using ZY1 02D VNIC data, taking the area around Zongjia Town, Dulan County, Qinghai Province, China as the research area. New hyperspectral image classification algorithms of GA-RF, PSO-RF, GNDO-RF, ASO-RF and MPA-RF were proposed, combining the traditional feature optimization algorithms such as GA algorithm, Particle Swarm Optimization (PSO) and Generalized Normal Distribution Optimization (GNDO) and the new feature optimization algorithms such as Generalized Normal Distribution Optimization (GNDO), Atom Search Algorithm (ASO) and Marine Predators Algorithm (MPA) with Random Forest (RF) classification algorithm. The proposed GA-RF, PSO-RF, GNDO-RF, ASO-RF and MPA-RF methods were applied to the vegetation type classification of the high-dimensional spectral-texture feature space. The results show that the texture images extracted with different window sizes and window movement directions have different advantages and disadvantages for vegetation type classification, and the inclusion of multi-scale texture features improves the overall classification accuracy of the data by 8.02% compared to the AHSI data. The proposed GA-RF, PSO-RF, GNDO-RF, ASO-RF and MPA-RF methods successfully improve the classification accuracy of the vegetation in the study area, and the overall classification accuracy (OA) is improved by 1.3234% to 2.3987%, compared with the traditional RF method. Of these, the MPA-RF method achieves the highest OA (88.9165％) and Kappa (0.8623). This study provides ideas for feature extraction, feature optimization and classification algorithm selection in hyperspectral remote sensing vegetation classification studies.