Abstract: Agricultural operations (such as tillage, soil turning, and deep loosening) significantly affect the soil structure, aeration, and nutrient distribution in crop production. The performance of soil-engaging components during operations can directly impact the agricultural machinery and the quality of soil preparation. Traditional mechanical analysis has been limited to capture the micro-movements of soil particles and the dynamic interactions in three-dimensional space. It is very necessary to accurately assess the tillage in the optimization of soil-engaging component. In particular, the SPH method can also offer a mesh-free approach. The complex behavior of soil can be expected to effectively simulate, including large deformations and interactions with mechanical components. In this study, the more accurate and comprehensive model was developed for the soil-engaging component interactions using SPH method. A three-dimensional contact algorithm was also incorporated using soil elastoplastic constitutive theory and penalty functions. The elastoplastic constitutive model was accurately represented the behavior of soil under various stress conditions, including elastic deformation and plastic yielding. The contact algorithm included the virtual planes and penalty functions. There was the complex contact behavior between soil particles and soil-engaging components. The algorithm was implemented in C++ on the CUDA (Compute Unified Device Architecture) parallel computing platform, thus leveraging the power of GPU (Graphics Processing Unit) acceleration. Numerical simulation was conducted using Visual Studio 2019. A series of experiments were also conducted to validate the accuracy of the improved model. The sand collapse experiment was firstly conducted to create a controlled environment, where the sand was allowed to collapse under its own weight. The diameter of collapse extension was theoretically calculated using empirical formulas. A comparison was also made on the simulation and the observed one. Theoretical diameter of collapse extension was 0.648 m with a relative error of only 2.16%. There was the close match between the simulation and experiment. The improved model was accurately predicted the soil behavior under large deformations. After that, the small-scale soil cutting apparatus was designed and then fabricated in the sand collapse experiment. The cutting forces were precisely measured under various conditions, including different tillage depths, speeds, and soil-engaging component angles. The predictive accuracy of the improved model was assessed to compare the numerical simulations with the experimental data. The model was further validated to predict the cutting forces for the soil cutting morphology. There was the high degree of correlation between the simulation and experiment. The patterns of sand cutting resistance were observed in numerical simulations and physical experiments, when the soil-engaging component. The combination of initial parameters was set as the depth of 0.06 m, a cutting speed of 0.01 m/s, and a penetration angle of 60°. The soil cutting forces were accurately predicted to represent the behavior of soil during cutting. The relative error was found to be only 7.02% after prediction, which was within an acceptable range for practical applications. Lastly, the systematic investigation was also implemented to explore the impact of various tillage depths (0.04, 0.06, and 0.08 m), tillage speeds (0.01, 0.02, and 0.03 m/s), and soil contact angles (45° and 60°) on the cutting force exerted by soil-engaging components. The tillage depth was the predominant influencing factor on the cutting force of soil-engaging components. There was the significant increase in the cutting force, as the tillage depth increased. There was the relatively minor influence of tillage speed and soil contact angle on the cutting force. This research can offer a solid scientific and technical support to optimize the agricultural machinery and tillage efficiency.