A quantitative analysis technique based on laser induced breakdown spectroscopy (LIBS) of neuro-genetic model is proposed. A three-layer back-propagation (BP) artificial neural networks (ANN) is constructed as a basic calibration model for LIBS analysis. The weight and threshold of the ANN are optimized by genetic algorithm. By combining calibration model with LIBS technique, high precision detection is achieved. The concentrations of Ba and Ni in soil samples are detected by using the given quantitative analysis technique. The mean relative errors are 4.15% and 6.06% respectively, and the correlation coefficients are 0.983 and 0.990 respectively. The presented results demonstrate that the neuro-genetic approach performs better than BP-ANN and conventional calibration method in LIBS quantitative analysis. The analytical results based on neuro-genetic approach in this study are well predicted, which provide a new modeling of high accuracy quantitative elemental analysis for LIBS technique.