The hash function is used as a one-way cryptography method for digital signature and message authentication. Hash values are provided using a mathematical and logical process, so they are different from the generators of random numbers. The position analysis of bits in plaintext and its hash is very suitable to show their relationship. The focus of this paper is to point to the best relations between the plaintext and hash bits, in which the difference between hash methods will be proven. In this work, we use distance correlation (dCorr) as a measurement function of precision statistical dependency between two vectors. The genetic algorithm (GA) is used to find a set of optimal positions between plaintext and its hash data with maximum dCorr. The results of the experiment regarding dCorr indicate that MD5 as a message digest method is different from the random function. Also, the proposed method compared with Tabu search (TS) and Simulated Annealing (SA) algorithms has a lower average execution time for 1000 pairs of plaintext and hash data.