As an important part of terrestrial ecosystem, forest is concerned by its huge carbon storage and carbon sequestration capacity. With the successful launch of China’s high score 1 (GF-1) satellite, it is possible to use NDVI data to realize the quantitative extraction of forest cover. However, due to the impact of rainy weather, operating costs and other factors, it is difficult to form NDVI GF-1 time series data, which cannot meet the demand for high precision extraction of forest cover. With the aim to solve this problem, Songshan was taken as part of the Henan GF-1/WFV NDVI and MODIS NDVI experimentation area, application of STAVFM algorithm was integrated, and GF-1/WFV NDVI time series data was used to generate the 8 d step, then from the time series data in NDVI feature extraction and spectral features were combined with GF-1/WFV. Finally, SVM classification method was used to realize quantitative forest coverage extraction. The research results showed that the NDVI GF-1/WFV sequence data generated by the STAVFM algorithm was ideal, which can solve the problem of the NDVI GF-1 time series data. The overall classification accuracy based on the SVM classification was 94.72%, which was improved by 4.90 percentage points compared with the classification results of the original GF-1/WFV image without fusing the characters of NDVI. This method provided a new way for high precision extraction of forest cover.