Heilongjiang Province is the largest commodity grain production base in China. Its crop information is of great significance for grain estimating, disaster prevention and reduction, and food security. This study explored the method of the planting information extraction to base on the hyperspectral data to help get plant distribution faster and better. Multi-spectral data have some limitation in distinguishing the planting information of corn and soybean,so in this paper we chose hyperspectral data to study the method of extracting main dry crop types in Qiqihaer City of Heilongjiang Province. After data preprocessing and optimum band selection, the crop types of corn, soybean, rice and other crops were extracted using the decision tree method. The results showed that the classification accuracy of each crop was more than 80% and its overall accuracy reached up to 88.71%. The classification results showed that the main land cover type was cultivated land in Qiqihaer City, which accounted for 70% of the city. And the plant area of corn was the main land cover type.