The car body consists of many parts, between which there are cavities and gaps that water, dust, and noise can enter. To prevent corrosion and reduce noise, these gaps are sealed with a paste featuring complex non-Newtonian properties. Sealing also serves a cosmetic function for visible areas, which demands high quality for better customer satisfaction. Usually, bead length can reach several meters with a height of 0.5-5 mm and a width of 5-40 mm. In this situation, optimizing the robotic paths and sealant flow can speed up production and reduce costs. Accurate and fast CFD modeling helps with planning the sealing process, shortening vehicle development cycles and minimizing costs. Due to the complexity of vehicle body geometry, arbitrary robotic movements, sealing bead length, free surface, and the complex rheological material properties, traditional CFD simulations have difficulties in modelling this process. This paper presents a new framework for modelling the sealing processes using CFD solver that is based on a novel meshfree Lagrangian Differencing Dynamics (LDD) approach, which considers the viscoplastic properties of the sealant material and the actual applicator with complex robotic movements. The LDD approach allows simplifying pre-processing and does not require the user to perform time-consuming and complex meshing. The use of the LDD approach in combination with GPU technologies will make it possible to conduct high-precision simulation of the sealing process of real objects in a few hours on a standard computer. This article presents basic introduction of the sealing process, the fundamental formulations of the LDD method and the mathematical model, implementation details, validation, and, finally, simulation results on realistic gar geometry with practical complex robotic movements.