The problem of multiscale computer modeling the processes of supersonic cold gas-dynamic spraying (SCGDS) the nanoparticles onto a substrate is considered. The problem relevance is related to the development of technologies for the manufacture of micro- and nanostructures used in modern microelectronics. In this work, a multiscale approach is used. As its basic mathematical models, the quasi-gasdynamic equations system supplemented by Newton’s equations system, which describes the dynamics of individual particles, and the molecular dynamics method are used. To solve gas-dynamic equations, a classical grid approach based on the finite volume method is used. The particle method is implemented by integrating equations using the symmetric Adams scheme. To implement the molecular dynamics equations, the Verlet scheme is used. When combining these models, many problems arise related to matching the different scales, stability to small disturbances introduced by different scales, and the efficiency of calculations parallelization. This work attempts to study the above problems. Matching the different scales is realized using splitting the general algorithm by physical processes and scales and observing the conservatism principles at the level of the modeled medium macroparameters. Stability to small disturbances is ensured by matching the spatial and time steps of the applied numerical schemes. The efficiency of parallelization is ensured by a combination of the Schwartz domain decomposition method and the algorithms for computers dynamic load balancing. Testing and refinement of parameters and procedures for transition between different levels components of a SCGDS multiscale model confirms the correctness of the proposed approach.