This　article　introduces　a　general　class　of　biased　estimator,　namely　a　generalized　diagonal　ridge-type　(GDR)　estimator,　for　the　linear　regression　model　when　multicollinearity　occurs.　The　estimator　represents　different　kinds　of　biased　estimators　when　different　parameters　are　obtained.　Some　properties　of　this　estimator　are　discussed　and　an　iterative　procedure　is　provided　for　selecting　the　parameters.　A　Monte　Carlo　simulation　study　and　an　application　show　that　the　GDR　estimator　performs　much　better　than　the　ordinary　least　squares　(OLS)　estimator　under　the　mean　square　error　(MSE)　criterion　when　severe　multicollinearity　is　present.