This paper presents an approach that combines the finite element (FE) modeling and genetic programming (GP) to provide accurate empirical stress intensity factor (SIF) equations for center-cracked steel plates repaired with adhesive-bonded double-sided fiber-reinforced polymer (FRP) patches. Several past studies in recent years independently showed that the reduction on the SIF of cracked structures after the patch repair is dependent on many factors such as bonding techniques, material parameters, geometric parameters, environmental factors. In this study, the SIF of the repaired cracked steel plate was considered to be a function of seven parameters including the crack length, elastic modulus of FRP material, shear modulus of adhesive material, dimensions (width, length, and thickness) of rectangular FRP patches, and thickness of adhesive layers. Empirical SIF equations were created by the data mining process of genetic programming analyses performed on a database created from the FE results. The SIF values obtained from these equations were also compared with an analytical equation to assess whether their ability to perform well on a certain design or not. It was found that the proposed SIF equations fitted well with the FE results as the squared Pearson correlation coefficients R 2 are higher than 0.9. In addition, the correlations between proposed equations and the analytical equation are approximately 0.8.