A self-consistent model of the pulsar magnetosphere is presented. The principal point of the model is its use of the massless approximation. In this approximation electrons and positrons are considered as massless particles, and the pulsar magnetosphere consists of regions of two types: (1) accelerating regions (A-regions) filled with particles moving with the speed of light [dynamical phase (DP)], (2) capture regions (C-regions) which, besides the DP, contain static phase (SP), i.e. captured particles resting at some coordinate system.