In order to avoid the disadvantages of genetic algorithms in solving numerical optimization problems, such as poor search ability and lack of diversity, an improved real-coded genetic algorithm(IRCGA) is proposed. The algorithm integrates two specially designed operators: the simulated binary jumping gene(SBJG) operator and the multi-directional crossover (MX)operator. The SBJG operates on chromosomes. It essentially simulates the insertion motion in binary jumping gene operation, that is, inserts the selected chromosomal block into the chromosomal locus in a random way, and realizes the translocation between the chromosomes within the population, which provids additional genetic diversity for the population. The MX operator expands the search area of the operator by increasing the crossover direction to improve the quality of the offspring and the search ability of the algorithm. The comparative experiment is carried out on the basis 11 examples. The results show that the improved operator can significantly improve the performance of the algorithm in solving numerical optimization problems. At the same time, compared with other advanced and effective algorithms, the IRCGA has strong search ability and can maintain a certain population diversity, thus verifying the effectiveness and feasibility of the improved algorithm.