This research suggests a genetic program (GP) which is equivalent in its work to Berlekamp – Massey algorithm to find the LFSR equivalent for a given chains Hence, the genetic program deals with a population of generating structures in a random state as independent programs that generate digital random bits chains ,and it is possible for the given chain to be one of them or part of them, and these programs work independently. Each a program that represents a suggested structure (recorded and described with a specific length and with a good link equality)in the population of random programs, given a value for fitness function that represents the digital value for the extent of fulfilling the results of the final population , for the description of the given chain. The function of the subordinate program which simulate with the genetic algorithm is to find the population of the final initial values for each a genetic program. The aim of this research is to build and deal with the Berlekamp – Massey algorithm throughout the genetic programming(GP) by following a way improves the situation of work of this algorithm in order to overcome some problems that it may face for example the disavaliability of the bits of the given chain, in addition to the probability of being non- linear generating chain.