Laser-induced breakdown spectroscopy (LIBS) has been considered a promising tool for agricultural analysis in the recent years due to its advantages over traditional techniques, such as fewer sample preparations and faster measurements. An adequate choice of equipment and optimization of parameters for measurements in soils becomes an important issue aiming at a better performance of the technique. The performance of two LIBS systems, using a monomode laser at 532 nm and a multimode laser at 1064 nm, were evaluated for soil analysis. The craters formed in the sample by action of the laser pulses studied by Scanning electron microscopy (SEM) aimed to analyze the efficiency of the lasers in the ablation process. The plasma temperature and electronic density, evaluated using the Saha-Boltzmann method, were similar for both lasers. The lower boundary of the Local thermodynamic equilibrium (LTE) condition was verified for both lasers. The signal-to-noise ratio (SNR) was analyzed for many emission lines of important elements of the soil, such as C, Fe, Mg, and K. Partial least square (PLS) analysis was used to verify the performance of both lasers for the quantitative analysis. From the PLS analysis, a better performance for the elemental quantification as found to 1064 nm laser and compared to 532 nm laser, i. e, higher coefficient R, a small Root Mean Square Error of Calibration (RMSEC) for C, K and Mg, and small Root Mean Square Error of Prediction (RMSEP) for K and Mg. In conclusion, the results suggesting the use of a multimode Laser at 1064 nm hence are more accessible, and less expensive, than a monomode Laser at 532 nm.