This paper examines how dietary knowledge and exercise knowledge are interacted each other in leading to healthier diet after hypertension diagnosis. It uses a sharp regression discontinuity design that exploits the exogenous cut-point of systolic blood pressure in the hypertension diagnosis. Using data from China, the results demonstrate that the influence of hypertension diagnosis on improving diet is significant only among people with good dietary knowledge; and within the people with good dietary knowledge, the influence is much larger among the people who do not believe the importance of exercise for health. The findings may imply that an equal emphasis on both diet and exercise might result in reducing the effect of dietary education on improving diet.