Simulations of nonequilibrium gas flows have garnered significant interest in modern engineering problems involving rarefied gas flow characteristics. Despite the popularity of the direct simulation Monte Carlo (DSMC) method in simulating such flows, its use in low-speed flows is limited by statistical noises. The information preservation (IP) method is a promising alternative known for its low noise properties. In this study, a new theoretical framework for the IP method based on kinetic theory is introduced to offer complete understanding for the transport properties of the preserved information. Specifically, we introduce a velocity-information joint distribution function (VIJDF) and derive its governing equation as well as the corresponding macroscopic transport equations. To ensure the accuracy of the IP method, the total stress/heat flux in IP, including information stress/heat flux generated during movement and collision steps and compensation stress/heat flux imposed in the compensation step, is matched to the molecular stress/heat flux in DSMC. To this end, a nonequilibrium model for the VIJDF is proposed to evaluate the compensation stress/heat flux. The parameters in the collision model of IP are theoretically determined by equating the transport coefficients associated with the preserved information to the coefficients of viscosity and thermal conductivity in DSMC. Numerical simulations for a variety of nonequilibrium gas flows, including low-speed Couette flow, Fourier flow, high-speed Couette flow, external force-driven Poiseuille flow, lid-driven cavity flow, and thermal creep flow, demonstrate that the IP method can achieve similar accuracy as the DSMC method with a much smaller sampling size.