In the current quality inspection of medical bags, there exists the problem of slow detection speed and low detection accuracy. Targeting at this problem, a new defect detection scheme was proposed based on HALCON in this paper. Firstly, manual positioning was executed based on different regional characteristics of medical bag patterns. Then the quick match of the target area was completed through the template matching algorithm based on shape. Finally, the defect detection was carried out with gray value subtraction matching algorithm. Detected defects were basically consistent with the actual ones, and the testing speed and the accuracy met the requirements. Pattern defects of medical bags can be detected quickly, accurately and effectively in this experiment which met the quality requirements of normal production line.