Abstract: In this paper we propose a new nonlinear stochastic dynamicevolution model for phase encoding in population of neuronal oscillatorswith different phases, and study the neural informationprocessing in cerebral cortex and dynamic evolution under the action ofdifferent stimulation signals. The evolution of the averagednumber density along with time in the space of three dimensions is described indifferent clusters of neuronal oscillator firing action potential atdifferent phase spaces by means of numerical analysis. The resultsof numerical analysis show that the dynamic model proposed in this paper canbe used to describe the mechanism of neurodynamics of attention andmemory, and it is proved that only the neural dynamic model in ahigh-dimension space can adequatelydescribe dynamic characteristics of the neural population, and muchuseful neural information may be lost in the early models of stochasticdynamics for phase coding.