This　article　considers　the　problem　of　variable　selection　for　a　class　of　linear　regression　models　with　instrumental　variables.　We　focus　on　the　case　that　the　covariates　are　endogenous　variables,　and　some　auxiliary　instrumental　variables　are　available.　An　instrumental　variable　based　variable　selection　procedure　is　proposed　by　using　a　modified　smooth-threshold　estimating　equations　(SEE).　The　proposed　method　can　attenuate　the　effect　of　endogeneity　of　covariates,　and　can　avoid　the　convex　optimization　problem.　Hence,　it　is　flexible　and　easy　to　implement.　Simulation　results　indicate　that　the　proposed　variable　selection　method　is　workable.