Chess is a well-known game. It has a huge searh area for each move. It had been used as one of many difficult challenges for Artificial Intelligence (AI) benchmarking. Human player has found many variations in white king, white rook, black king (kRK) chess endgame. This paper discusses the implementation of genetic programming (GP) algorithm into the KRK (white king,white rok,black king) endgame, combining a set of elementary strategies that are generally done by a chess player. The influence of GP parameters is observed to individual mean fitness value, individual best fitness value and how much time is needed for GP processing. There are six parameters,each of them were examined for eight times. The parameters included are the number of individual in population, the number of generation, crossover and mutation probability, the number of depth, and the a and b multiplying factors. The result from 48 imes trials, 43 ended with draw and only five tests ended wit winning. It indicated that GP parameters influenced the value for individual mean fitness and individual best fitness, while parameters that influence the GP parameters processing time only the number of individual, the number of generation and the number of dept. Although the GP program has not been proven to be effective enough when it is played against another AI program (using minimax algorithm) or even againts human player, the playing stategies proposed by GP are relatively understandable by human. Keywords: Chess Endgame, King-Rook-King, Genetic Programming,Artificial Intelligence