This　study　focuses　primarily　on　sulfur　dioxide　(SO2)　emissions　control　problem　in　a　wet　flue　gas　desulfurization　(WFGD)　process,　and　our　objective　is　to　design　an　intelligent　control　system　so　that　the　outlet　SO2　concentration　satisfies　the　SO2　emission　standard.　In　our　approach,　a　multimodel　control　framework,　which　is　made　up　of　a　linear　robust　controller　and　a　neural　controller,　is　integrated　with　the　invasive　weed　optimization　(IWO)　algorithm　in　an　elegant　fashion　and　used　for　SO2　emissions　control　purposes.　A　case　study　is　carried　out　based　on　operation　data　from　a　600　MW　coal-fired　unit,　and　simulation　results　show　that　IWO-based　automatic　clustering　can　identify　different　operating　modes　in　the　WFGD　process　with　high　accuracy.　Further,　the　established　multimodel　control　system　can　remove　SO2　emissions　effectively.　Experimental　results　show　that　SO2　emissions　can　be　removed　effectively　with　the　proposed　method,　and　this　could　provide　engineering　guidance　to　design　a　WFGD　control　system.