Agoodmodelcanextractusefulinformationaboutthetarget'sstatefromobservationseffectively.Therearemanymodelsusedtotrackinga,maneuveringtargetsuchasconstant-velocity(CV)model,Singeraccelerationmodel(zero-meanfirst-orderMarkovmodel)andcurrentmodel(mean-adaptiveaccelerationmodel),etc.Whileduetothecomplexityofmaneuveringtarget,toseekthetargetmodelwhichcangetbetterperformanceisstillasubjectworthyofstudy.BasedonstatisticsrelationbetweentheautocorrelationfunctionandthecovarianceofMarkovrandomprocessing,thispaperdevelopsamodelwhichcanadaptivelyadjustsystemparametersonline.Simulationsshowthegoodestimationperformancegetbythemodeldevelopedhere,andcomparingCV,Singerandcurrentmodels,themodelcanadaptivelygetthemodelparameterwhiletrackingthetrajectoryandneedn'tdoingseveralteststoobtainaprioriparameter.