Abstarct: Developing the industry and extensive demands for energy and mineral resources, such as hydrocarbon reservoirs, the desire for deeper excavation to access the oil and gas reservoirs and rich mines in the earth's crust is expanding. Consistency in the production of these resources requires the preparation of the stable opening that keep the shape during time. The drilling experience in different parts of the crust show that the lack of awareness of applied stresses in the earth's crust is the most important factor in the incorrect design of drilling and as a result the collapse of these openings. Many methods have been introduced and investigated by researchers to measure stresses in recent years, which mostly are expensive and sometimes impossible due to operational conditions. Monitoring the displacements and deformation of the wellbore has been of interest to researchers since the past. Nowadays, a was variety of laboratory investigations have been carried out to assess the behavior of borehole breakouts caused by the inhomogeneity of the stress state. Nevertheless, no reliable method has been yet provided to establish a relationship between the breakout and the stress magnitudes. In this study, a method baxsed on back analysis was used to estimate the stress magnitudes. First, using the numerical analysis baxsed on Distinct Element methods, a model has been developed and its results have been used to train the forward model structured by the relevant vector regression. Then, harmony search optimization was coupled to inverse the proposed model, and finally the backward model has been generated. To verify the proposed model in a real field case, an algorithm has been developed to detect the pattern of the breakouts from ultrasonic imaging tool. The results have been composed with the combination of geomechanical modeling of the reservoir, which has done with the help of well survey data, the regression models, and then the ratio of the measured field stresses was calculated. Finally, in a well, the calculated stress values should be compared with the results of the field values of the Leakage-off test. The result shows that compared to the field data of the Leak-off test, the previously developed method can be replaced with the proposed method.