Abstract: Systematic errors have recently been founded to be distinctin the zonal mean component forecasts,which account for a large portion of the total monthly-mean forecasterrors.To overcome the difficulty of numerical model,the monthly pentad-mean nonlinear dynamic regional prediction models of the zonal mean geopotential heightat 200,300,500 and 700 hPa based on a large number of historical data(NCEP/NCAR reanalysis data) were constituted by employing the local approximation of the phase space reconstruction theory and nonlinear spatio-temporal series prediction method.the 12-month forecastexperiments of 1996 indicated that the results of the nonlinear model are better than those of persistent,climatic prediction and T42L9 model either over the high-and mid-latitude areas of the northern and southern hemisphere or over the tropical area.the monthly-mean heightroo-tmean-square of T42L9 model was considerably decreased with a change of 30.4%,26.6%, 82.6% and 39.4% respectively over the high-and mid-latitude of the Northern Hemisphere,over the high-and mid-latitude of the Southern Hemisphere,over the tropics and over the Globe respectively,the corresponding anormaly correlation coefficients over the four areas were respectively increased from 0.306 to 0.312,from 0.304 to 0.429,0,739 to 0.746 and from 0.360 to 0.400(averagely a relative change of 11.0% over the Globe) by nonlinear correction after integration.Itimplys that the forecasts given by nonlinear model includes more useful information than those of T42L9 model.