(1.College of Earth Sciences, Yangtze University, Wuhan 430100, China): This research is based on the Google Earth Engine (GEE) remote sensing data cloud computing platform, using the 1443 scene Landsat Surface Reflectance long-term remote sensing image data, using normalized vegetation index, pixel Methods such as dimidiate pixel model, coefficient of variation, linear regression analysis and CA-Markov model monitor the dynamic changes of vegetation coverage in Qinhuangdao from 2001 to 2020, and predict the vegetation coverage in 2025. Research shows that: (1) From 2001 to 2020, the vegetation coverage of Qinhuangdao in the past 20 years was generally good, with medium to high vegetation coverage and high vegetation coverage as the main factors, and the sum of the two areas reached more than 65% of the total area; The sum of the area of ??low vegetation coverage and medium and low vegetation coverage does not exceed 20%. (2) The vegetation coverage around rivers, lakes and urban areas fluctuates greatly, and the coefficient of variation ranges from 0.8 to 3.2. (3) The area where the vegetation coverage of Qinhuangdao has changed significantly in the past 20 years accounted for 57.7%, of which the decreased area accounted for 17.7%, the increased area accounted for 40%, and the area where the vegetation coverage remained basically unchanged accounted for 42.3%, and human factors remained unchanged. It is the main factor affecting the change of vegetation coverage. (4) The forecast results show that in 2025, Qinhuangdao will still be dominated by high vegetation coverage, accounting for 47.4%, an increase of 0.9%, and low vegetation coverage and medium vegetation coverage will decrease by 0.8% and 0.4%, respectively. Both low vegetation coverage and medium-high vegetation coverage increased by 0.1%. This research can provide data reference and theoretical basis for the environmental governance and urban planning of Qinhuangdao.