The main activity to verify software quality and reliability is software testing. To obtain the correct performance of the software execution environment, software testing is essential. Software testing is an expensive and time consuming process so much effort has been spent to automate it. One of the software testing techniques is test data generation. Test data generation in program testing is the process of identifying a set of test data that meets the test criteria. To solve the problem of test data generation, it is converted to an optimization problem so that we can take advantage of optimization techniques to solve this problem. Metaheuristic search techniques are successive approaches in this area. The use of metaheuristic search techniques for the automatic generation of test data has been a burgeoning interest for many researchers in recent years. These techniques offer much promise in regard to these problems. Metaheuristic search techniques are high-level frameworks, which utilize heuristics to seek solutions for combinatorial problems at a reasonable computational cost. Evolutionary algorithms are one of the most common methods of searching and nowadays are used in many real-world problems. One of the applications of evolutionary algorithms is generating test data automatically. This thesis aims to generate test data using evolutionary algorithms and considers branch coverage as quality criteria. The search technique that has been used in this thesis is learnable evolution model (LEM). LEM employs machine learning to generate new populations. Speci?cally, in Machine Learning mode, a learning system seeks reasons why certain individuals in a population are superior to others in performing a designated reasons, expressed as inductive hypotheses, are used to generate new populations . Furthermore, the knowledge gained from the study of earlier stages is used to enhance the speed. So LEM is used in order to take advantage of the knowledge in the form of hypotheses. In other words, the idea is that the search space is reduced to search faster. As mentioned, the problem of test data generation is an optimization problem. Before solving the optimization problem, no information about the locations of the global optimum is available . Thereby, orthogonal arrays have been used to generate an initial population of points that are scattered uniformly over the feasible solution space , so that the algorithm can evenly scan the feasible solution space once to locate good points for further exploration in subsequent iterations. The test functions in similar studies have been chosen to evaluate the performance of the proposed method. The proposed method is used to generate fewer test data to achieve maximum branch coverage which experiments is provided this fact. In order to evaluate performance of the proposed method, the results have been compared to the other methods' outputs, which show that the proposed method has a relatively better performance comparing the other approaches. Keywords : Test Data Generation, Metaheuristic Search Techniques, Learnable Evolution Model