Aiming at the problems such as single evaluation index and significant deviation of horizontal comparison of subsoiling operation quality, by loading intelligent terminal equipment on agricultural machinery, collecting GNSS, operation depth and other information of agricultural machinery operation, analyzing the spatial trajectory data of agricultural machinery, calculating the feature vectors of agricultural machinery deep loosening operation, such as operation area, operation efficiency, operation coverage, operation depth compliance rate, operation time utilization rate and so on, based on the minimum-maximum normalization method, the characteristic vectors were normalized, and the weight of each feature vector was determined by the expert scoring method. The agricultural machinery deep loosening operation quality index (AMQI) was further calculated to evaluate the comprehensive quality of agricultural machinery deep loosening operation. In order to verify the quality index of agricultural machinery subsoiling operation, from January 1st, 2022 to December 31st, 2022, the agricultural machinery experiment was carried out in Hongxing Farm of Heilongjiang Province. The vehicle-mounted intelligent terminal was installed on 33 tractors on the farm, and the subsoiling operation data was continuously collected to calculate the quality index of agricultural machinery subsoiling operation. The results showed that the normalized feature vector eliminated the dimensional influence between different features and reflected the size distribution of agricultural machinery in features. The value of AMQI also reflected the comprehensive quality of deep loosening operation of agricultural machinery, compared with the standard area evaluation method, the rating dimension was more, and the evaluation was more comprehensive and scientific, which can provide data support for horizontal and vertical evaluation of agricultural machinery operation.