This paper aims to meet the new requirements of modern analytical and testing technology development, and to promote the application of Laser-Induced Breakdown Spectroscopy (LIBS) in the field of element analysis, especially for the measurement of rare earth element in soil. A LIBS system combined with calibration curve method was used to quantitatively analyze samarium (Sm) in the soil of Bayan Obo rare earth mining region. Firstly, the samples containing 1%, 5%, 10% and 20% Sm₂O₃ were prepared by Standard Addition Method (SAM) with the soil of national standard material GBW07402a as the base. Secondly, through analyzing the substrate excited by different laser pulse energy parameters, the influence of laser pulse energy parameters on the spectral line intensity and Signal to Back Ratio (SBR) was researched, an optimum laser pulse energy parameter was finally selected for the next measurement. Thirdly, in order to get and study the linearity of the calibration curve constructed between the peak area and the Sm concentration, the original spectra data were processed with multiple peak Lorentz fitting method without background subtraction (MFM) and Concatenation-based Integration Method (CIM) with background retention, respectively. Finally, according to the calibration curve, the concentration prediction was carried out, and the detection performance of LIBS for Sm in soil samples of rare earth mining area was preliminarily evaluated. The results show that the matrix effect of the soil can significantly broaden the emission lines of Sm element, which makes it impossible to distinguish them from each other. However, the effect of the soil matrix on sodium (Na), potassium (K), Titanium (Ti) and iron (Fe) is much weaker than that on Sm. By comparing the spectral region of interest, the 410 nm-band and 470.44 nm emission lines were identified and selected as the analysis lines, and subsequently used for quantitatively analysis. Results show that calibration curves for Sm element constructed by the peak area and concentration have good linear correlations and most of the linear relationships of the regression coefficients (R²) for the Sm emission lines are better than 0.99. Compared with the results by using MFM, CIM could obtain better linear correlation, and the maximum of was 0.99927 for the 410 nm-band. The better analytical predictive skill of LIBS measurement by using the leave-one-out method with CIM data was found as well, the relative errors of the prediction for both the analysis lines were all within 1% for the 3# sample with the Sm concentration of 4.310%. The achievements of this study demonstrate that the LIBS spectral analysis is capable of monitoring special elements in the rare earth mineral sample, which meets new requirements of modern analysis technology, and provides an experimental basis for the development of portable rare earth element detector as well.