Based on the characteristic that DJI Phantom 4 multispectral version can obtain surface optical photos and multispectral images simultaneously, this paper introduces a method, using photogrammetry technology to construct a canopy height model(CHM), and combines multispectral images to apply mangrove classification. The above method was applied to the research area of Gaoqiao Town, Zhanjiang Mangrove National Nature Reserve, Guangdong Province. The results show that the combination of mangrove and it´s CHM constructed based on UAV photogrammetry technology and multispectral imagery can greatly improve the classification accuracy of mangrove, which can replace the traditional airborne lidar data greatly; the DJI Phantom 4 multispectral version has high data acquisition efficiency, and can achieve high-precision mangrove classification, and has a very good application prospects in mangrove monitoring field.