Tosolvetheproblemofthemismatcherrorsbetweentheactualandpresumedsignalsteeringvectorsinairborneradarspace-timeadaptiveprocessing(STAP),arobustreduced-dimensionSTAPmethodbasedonbi-iterativealgorithm(BIA)andsecond-orderconeprogramming(SOCP)isproposed.FirstlytheweightvectorisdecomposedintotheKroneckerproductofthespatialandtemporalweightvectors.Thenbyimposingaboundontheerrorsbetweentheactualspatialandtemporalsteeringvectorsandthepresumedsteeringvectors,respectively,theoptimizationoftheworstcaseperformanceisexecutedandtransformedintotheSOCPformulation.FurthertheBIAisutilizedtoresolvethetwolowdimensionweightvectors,andfinallythefulldimensionSTAPweightvectorissynthesized.TheproposedmethodcansignificantlydecreasethetrainingsamplerequirementandthecomputationalcomplexitywhilemaintainingtherobustSTAPperformance,whichismorevaluableforpracticalengineeringapplication.Simulationresultsverifytheeffectivenessoftheproposedmethod.