In clinical practice, Myocardial Blush Grade (MBG) has been used to obtain information about microvascular condition in myocardial infarction by using coronary angiogram. Quantitative Blush Evaluator (QuBE) program was developed for the calculation of myocardial perfusion score. Calculation of QuBE values is often affected by patient motion and become inaccurate. In this paper, we proposed an algorithm to reduce undesired motion in coronary angiogram. This algorithm correct frame motion by shifting each single frame according to the best correlation with the first frame. The effectiveness of this stablizing method achieved by searching scale-invariant feature from each frame of coronary angiogram in order to find the best correlation between two frame. The results showed that MBG categorization based on modified QuBE program exactly match with the original QuBE program. In addition, results also showed that application of stabilization algorithm using SIFT method decreased the deviation by 15% therefore it increased the accuracy of QuBE value calculation. Finally, this new algorithm also decreased the execution time by 71% so the doctor could faster patient diagnosis. In conclusion the new algorithm could enhance the qualitiy of QuBE value calculations in MBG scoring for coronary angiogram. Keyword: coronary angiography, myocardial blush grade, stabilization, scale-invariant feature transform, quantitative blush evaluator.