The silylation of methylphosphonic acid(MPA) was studied based on the artificial neural network(ANN) technique in three organic solvent systems,including pyridine-toluene,acetonitrile-toluene and tetrahydrofuran(THF) system.The influences of temperature,time,derivatization reagent and its amount on the silylation derivatization efficiency were investigated and optimized.The type of derivatization reagent and its amount were confirmed as the main influencing factors.On the basis of the results obtained,trimethylsilyl(TMS) derivatization and t-butyldimethylsilyl(TBDMS) derivatization methods for MPA,isopropylphosphonic acid(IPA) and pinacolyl methylphosphonic acid(PMPA) were presented,respectively.10%(by volume) of N,O-bis(trimethylsily)trifluoroacetamide(BSTFA) and N-methyl-N-(t-butyldimethylsilyl)trifluoroacetamide(MTBSTFA) were used as derivatization reagent,respectively,in three different organic systems.The results showed that TBDMS derivatives of three compounds were more stable than their TMS derivatives.The yields efficiencys of TBDMS derivatization of three alkylphosphonic acids were evaluated by using 31P{1H}-NMR spectroscopy,indicating that the derivatization reaction was nearly complete.