Abstarct: Determining the depth and geometry of the baxsement in sedimentary basins is one of the strategic goals of many exploration projects, especially groundwater and hydrocarbon reserves. Measurement and modeling of potential field data are one of the most widely used geophysical methods in this field. The main purpose of this dissertation is to investigate the possibility of using public search algorithms, ant colonies, and colonial competition as an alternative to local response search methods such as Marquardt-Levenberg and Gauss-Newton; in nonlinear modeling, the thickness of sediments in sedimentary basins using gravimetric and magnetometric data. For this purpose, first by examining the existing leading models to calculate the model response at data measurement points, the leading models (Telford-1976) for modeling, gravimetric data, and the model (Kunartnam-1981) for the model The construction of magnetometric data was considered. To perform inverse modeling of gravimetric and magnetometric data by the proposed algorithms and also to present the results visually, the required codes were written in the MATLAB software package. In order to validate the developed codes, first, designing and implementation of the mentioned algorithms for the data belonging to two synthetic and magnetic and gravity models were validated in two stages of noise-free data. A quantitative comparison of the difference between the data obtained from the initial models and the calculated data was performed by determining the root mean square of the error, which is a parameter for gravimetric modeling from 0.23 mGal, and For magnetometry modeling, it did not exceed 7.1 nT, which shows the good performance of the designed algorithms on both types of modeled data. Also, in this study, the effectiveness of this method against the amount of possible common noises on the mentioned synthetic data was investigated. The results of this study; Demonstrates the good stability of this algorithm for modeling data of both types of artificial models against Gaussian white noises. By adding up to 8% white noise to artificially expensive anomaly data, and up to 10% white noise to artificial magnetic anomaly data, both algorithms have good stability and were able to model Rebuild the default with proper care. Also, the performance of the colonial competition algorithm on the data obtained from an artificial model of a deep sedimentary basin with density with parabolic changes to the depth, baxsed on the leading model (Gallardo-2003) on two types of data without noise and with white noise quantitatively baxsed on the root mean square error, it was estimated that the results show the proper performance of the algorithm for this type of sedimentary basins. To evaluate the performance of the developed codes, field data from the case study of gravimetric and magnetometry data of Moghan sedimentary basin in northwestern Iran were used. Comparison of modeling results with the results of existing geological studies as well as the results of past geophysical studies through seismography and gravimetric modeling using the Marquardt-Levenberg method; Demonstrates proper performance of written code. Finally, due to the differences in the density distribution in sedimentary basins; Inversion of Field Data Gravimetric studies were performed on two Los Angeles sedimentary basins in California and the Atacama Desert sedimentary basin in northern Chile using a colonial competition algorithm. In modeling the sediment thickness of Atacama Desert sediments, the forward model with constant density contrast was used and in the case of Los Angeles Basin, the forward model with variable specific gravity was used Quadratic. The models obtained in each of these two modeling’s are in agreeing well with the results of previous gravity and seismic studies conducted in those two basins.