This paper suggests an effective method for facial recognition using fuzzy theory and Shannon entropy. Combination of fuzzy theory and Shannon entropy eliminates the complication of other methods. Shannon entropy calculates the ratio of an element between faces, and fuzzy theory calculates the member ship of the entropy with 1. More details will be mentioned in Section 3. The learning performance is better than others as it is very simple, and only need two data per learning. By using factors that don't usually change during the life, the method will have a high accuracy.