Additive manufacturing technology is often used to prepare complex metal parts due to its high processing speed, high precision, and there is no need for molding for shaping. The preparation of component gradient samples is more popular in the manufacture of metal additives, Since the technology is not yet mature. There are often many defects in the workpiece. It is very important for the quality monitoring of additive manufacturing products to make the study of component distribution characterization method, which is suitable for additive manufacturing samples. The macroscopic component distribution characterization methods are mainly composed of Laser-induced breakdown spectroscopy combined with original position statistical distribution analysis (LIBS-OPA) and Spark source atomic emission spectroscopy combined with original position statistical distribution analysis (Spark-OPA), due to the large excitation spot, Spark- OPA is not suitable for layer-by-layer analysis of additive manufacturing samples. LIBS-OPA has gradually been used to characterize the element distribution of metal block samples with the advantages of multi-element synchronous positioning analysis, high spatial resolution, large optional analysis area, small sample damage, This method can achieve high-precision composition distribution characterization of metal workpieces. In this paper, the composition distribution of gradient stainless steel samples prepared by additive manufacturing technology was studied by laser induced breakdown spectroscopy. By optimizing the instrument parameters and analysis conditions, the analytical sensitivity and signal stability was ensured. The optimum test conditions were selected as follows: laser lamp voltage 1.32 kV, Q-switching delay 280 μs, sample chamber argon pressure 6 300 Pa, spot diameter 200 μm, 0 pre-denudation, integration of 15 denudations. Under this condition, the calibration curves of Cr with spectrum line of 298.9 nm, Ni with spectrum line of 218.5 nm, Mo with spectrum line of 203.8 nm, Si with spectrum line of 212.4 nm, P with spectrum line of 178.3 nm, C with spectrum line of 193.1 nm, Co with spectrum line of 384.5 nm, and Mn with a spectrum line of 293.3 nm was plotted. Most element determination coefficients exceed 0.99, Two gradient stainless steel samples prepared with different multi-pass powder feeding processes were scanned by LIBS-OPA. The distribution information of eight elements in the deposition surface of the samples was obtained. The quantitative distribution results had good agreement with Spark-OPA, and the quantitative accuracy has been verified by spark direct reading spectrometer. The study achieved a layer-by-layer analysis of additive manufacturing samples and selected the sample preparation process by composition distribution results. At the same time, the causes of cracks in the samples were also discussed through the characterization results of composition distribution. This research can play a guiding role in the improvement and perfection of the manufacturing process.