Traditional methods of agricultural disaster monitoring not only lag information, but also cause inaccurate damage determination due to manual survey methods. Therefore, timely and accurate access to agricultural disaster information is of great significance for disaster relief, agricultural insurance claims, optimization of planting structure adjustment, and precision poverty alleviation. This paper proposes a remote sensing extraction method for agricultural disaster information based on the temporal and spatial statistical characteristics of vegetation index. This method can quickly and accurately extract agricultural disaster information. This method uses the two Meng cities in the Inner Mongolia Autonomous Region as the research area, extracts the agricultural disaster information of the region in 2019, and calculates the disaster area and level of the Meng city in the region. The long sequence of agricultural disaster information extraction avoids the problems of lagging and inaccurate information of traditional agricultural disaster monitoring methods. In the experiment, the MODIS reflectivity products from late June to late September were used to distinguish the geographical location and the type of cultivated land, and the information on agricultural disasters in the key growth period of the crop in 2019 in the study area was obtained. The total affected area reached about 4.3 million mu. The method of this paper is used to extract agricultural disaster information. The overall accuracy Pc is 0.97 and the Kappa coefficient is 0.59. Compared with the traditional method, the accuracy of the extraction results is improved by nearly 15%, which indicates that the agricultural disaster information based on the vegetation index space and time is proposed in this paper. The remote sensing extraction method is suitable for agricultural disaster monitoring with wide space and long time series.