According to the technical demand of stored grain pests forecast in China,research status and open questions about stored grain pests forecast was analyzed. In order to solve the problem, the increasing population of insects, growth rhythm, and chaotic characteristics of time sequence were discussed. The phase space reconstruction,Poincare section, and the return map and auto correlation function of chaos theory are applied tentatively to analysis and predict time series of pests in non-stationary stage.Chaotic theory is applied to the field of prediction for stored grain insects to seek the innovation and break through point of stored grain pest’s prediction technique, which provided technology support for stored grain pests prediction system and had guiding significance for the improvement of stored grain pest’s forecasting technology and food security.