Abstract: Automatic detection of granary storage quantity with fast speed and high accuracy is an important technology for national food security. In this paper, according to the characteristics of grain peaks in different granary type (flat warehouse, squat silo and vertical silo), the mathematical relationship between granary storage quantity and the bottom/side pressure of granaries is established, a new on line model for the grain storage quantity measurement is also proposed. For the given grain and storage type, the storage quantity is only related to the bottom pressure and side pressure on the granary and the grain height in the granary, which make it possible to measure the storage quantity by pressure sensors, as long as the above pressure value and grain height is determined, the storage quantity can be obtained with high accuracy. Accord to the measurement requirement on the accuracy and robustness on the grain storage quantity, the pressure sensors are put with the shape of two circles, each circle includes multiple pressure sensors. This is because the inhomogeneous and randomness of the pressure value on the bottom and side of the granary. Therefore, more sensors can detect the pressure value on the bottom and side with high accuracy. However, the more pressure sensor used, the higher on the cost of the measurement system. Based on the tradeoff on the cost and accuracy, the two circles setup is proposed to measure the pressure value with high accuracy. The inner circle which is far from the side wall provides the pressure value on the granary bottom and the outer circle which is close to the side wall provides the pressure value on the granary side. In this method, the two circles setup not only eliminated the inhomogeneous and randomness of the granary pressure value measurement, but also reduced the number of the required pressure sensors. With the pressure values obtained, according to the principle model of the storage quantity measurement, this paper proposed a new model based on polynomial expansion to measure the storage quantity. The model use the polynomial of the average value on the inner pressure sensors to estimate the bottom pressure value, and use the polynomial of the average value on the outer pressure sensors to estimate the side pressure value, and obtain the grain height based on the above two polynomial. The proposed model can obtain the polynomial coefficient and order based on the sample of the pressure values automatically, which provides the high accuracy on the grain storage quantity measurement based on the pressure sensors. At the same time, the granary type such as granary shape and size are also introduced in the model, which improved practicality for different types of the granary of the proposed model. Real Practical storage quantity detection results show that the detection error of the proposed model is less than the requirement 2.5%, and the cost is reduced significantly, which demonstrates that the proposed granary storage quantity detection model is effective and well meet the demand of national granary storage quantity detection.