Abstract: In this study, compression strength and physical properties of the forty concrete carrot specimens taken from some buildings which collapsed by 1999 earthquakes were investigated and the correlations between compressive strength and physical properties determined. Compressive strength prediction models were established according to the physically properties of the specimens. These models tried to maximize using Genetic Algorithm (GA) and Linear Programming (LP) depending on specimens’ properties, ultrasound values, Schmidt hammer’s values and one-axis compressive strength. LP and GA were applied in order to maximize the carrot specimens’ compressive strength and optimum properties determined. Finally, it was seen that one-axis compressive strength of the forty carrot specimens were changing between 8-15 kPa. However, maximized one-axis compressive strength were calculated 14.97 kPa with LP and 15.27 kPa with GA.