Due to intensive heterogeneity and micro-nano-meter-scale pores in tight formation, flow mechanism differs significantly from that in conventional reservoirs. Therefore, the capability to simulate pore scale flow in tight formation is of great importance in many applications, such as oil extraction from tight reservoirs and accurate prediction of oil production. In this paper, a 3D random network model which can characterize the heterogeneity of tight formation is proposed. Based on the established network model, a flow mathematical model in consideration of boundary layer effect is developed, whereas parameters in boundary layer thickness formula are determined by Particle Swarm Optimization algorithm. Then, repeated iterations are used to solve the flow mathematical model until the relative error of pressure in each pore between two adjacent iterations arrives at the given error, which guarantees the accuracy of the simulation result. The result is then validated by micro-tube experiment. Furthermore, factors influencing pore scale flow in tight formation are studied. Results show that: as a result of boundary layer effect, absolute permeability is no longer a fixed value, it increases with the increase of pressure gradient and reaches a stable value when displacement pressure gradient is large enough. In low displacement pressure gradient, non-linear flow appears. As average coordination number increases, connectivity of tight formation becomes better, absolute permeability and velocity are larger at the same pressure gradient; When pore radius remains unchanged, as aspect ratio increases, throat radius becomes smaller, effective space for fluid flow is compressed, leading to smaller absolute permeability and velocity at the same pressure gradient; Change of 3D random network model size through varying throat lengths has little effect on pore scale flow; With the increase of fluid viscosity, boundary layer becomes thicker and the effective flow space grows smaller, resulting in lower permeability and velocity at the same pressure gradient. When displacement pressure gradient is large enough, absolute permeability tends to a stable value, which has nothing to do with fluid viscosity.