This paper provides empirical evidence that emotions help to explain reference-dependent preference. I develop the reference-dependent model which includes sympathy, anticipation, personal angry, envy, angry to fairness. Some of my findings are :(1) the difference between the actual and expected influences on decision making. (2) the difference between the actual and expected take rate is a much better predictor of experienced emotions and destruction than the difference between the actual and the fair take rate. (3) shame than guilt gives important role on decision making. (4) reduce destruction in proportion as player’s sympathy. (5) the difference between the actual and expected explains reference dependent. (6) the difference between the actual and expected gets personal angry or envy. (7) the degree of personal angry relates the degree of punishment.