Under the condition of D-differential, the convex fuzzy programming and its Lagrangian dual problem are discussed based on saddle point theory. Then, a more reasonable and easy-to-use form of KKT conditions in convex fuzzy programming is obtained, and a detailed example is used to illustrate the feasibility of the proposed method. Moreover, the relationship between the optimal solution of convex fuzzy programming and the saddle point of Lagrangian function is given.