Abstract: Rice is one of the main food crops in China, and the detection of heavy metal content is of great significance for its safety and quality. Laser-induced breakdown spectroscopy (LIBS) is expected to overcome the shortcomings of traditional methods for detecting time-consuming and achieve rapid insitu quantitative detection of heavy metals in crop plants such as rice. The collinear double-pulse laser-induced breakdown spectroscopy (DP-LIBS) technique was used to analyze the heavy metal Cd elements in rice leaves. Because different characteristic spectral lines have different effects on the results, in order to obtain more accurate and stable analysis results, the multi-point ablation test method of the sample surface was used to discuss the influence of the three resonance lines of Cd I: 228.8 nm, Cd II: 214.4 nm, and Cd II: 226.5 nm on the coefficient and prediction results of the calibration curve. The comparative study found that the best spectral intensity can be obtained when the 532 nm laser is excited before the 1064 nm laser, and the two pulses have a time interval of 0.5 μs. Among the three analytical lines, the calibration coefficients R2 of Cd I: 228.8 nm, Cd II: 214.4 nm and Cd II: 226.5 nm spectral lines are 0.86, 0.60, and 0.93 respectively, and it is found that the quantitative results of atomic spectral lines are higher than that of the ionic spectral lines. Further study shows that the relative prediction errors of Cd I: 228.8 nm spectral lines are less than 10%, and the detection limit is 3.03 mg/kg. Experiments show that by optimizing the detection conditions of LIBS, the detection of heavy metal contents in rice leaves can be achieved. In addition, the optimized spectral acquisition and characteristic spectral line selection methods in the experiment are also expected to be applied to the detection of heavy metal components in other different agricultural products.