This study investigated the change of impact resistance in concrete with respect to their sample size and compressive strength as experimental and theoretical. In the experimental design phase, a three-parameter, four-level L16 (43) Taguchi orthogonal array was used. For the Charpy impact tests, 16 serial samples were prepared with the following characteristics: sample width/sample length (B/L) ratio of 0.20, 0.25, 0.40 and 0.50; a sample width (B) of 100, 150, 200 and 250 mm, and a relative notch length (a/B) of 0.1, 0.2, 0.3 and 0.4. Experimental data were analyzed by two different methods. In the first stage, the optimum values were determined by performing analyses with the Taguchi method, while in the second stage, a mathematical model created by the Response Surface Method (RSM) was used as the objective function in a genetic algorithm to obtain the optimum values. Validation tests were performed based on optimum values and the results were compared. As a result of the study, it was observed that the change in sample size and compressive strength significantly affected the impact resistance of the concrete. A mathematical model was created to determine the impact resistance of the concrete.