**Abstract:**
Graph partitioning is one of the most important problem in optimization and graph theory. This problem falls into NP-problems and does not have an exact solution. In this paper, we intend to propose a new solution to optimize balanced undirected graph partitioning problem. The solution is an evolutionary algorithm, so called Shuffle Frog Leaping. A fitness function will be used to evaluate the suitability of each frog in each iteration. In order to measure the performance of our method, we partitioned a random graph with 10000 nodes into arbitrary numbers. Furthermore, we compare our result with two well-known algorithms called Genetic algorithm and Artificial intelligent.

**Keywords:**
Balanced Graph partitioning, Shuffle Frog Leaping, Optimization algorithms.