Abstract: An on-line coupled regional climate and atmospheric chemical model system has been modified and used to compare the on-line and off-line methods for direct radiative forcing and climate response simulation of sulfate aerosol over China area under the same model system and the same polutant release inventory condition.The differences between sulfate column burden and TOA radiative forcing without climate response have been discussed based on four-month simulation,i.e.January,April,July and October in 2000.The results reveals the maximum difference of sulfate burden between online and off-line methods emerges in July,and the percentage ratio to on-line burden of the difference is less than 3% at the whole modeling area average scale,but there are about 13% to some regions.It has been concluded that the sulfate burden has less difference at large scale,but is obvious at regional scale.The radiative forcing at TOA without climate response has a variable range between-1.5 W/m2 and 1 W/m2 in the two methods,and positive value locates in the negative sulfate burden difference region.At the whole average scale,the difference of radiative forcing is about 11% to that of on-line result,but the difference is more than 30% to that of on-line result in the central and northern region.At the regional scale,the radiative forcing has been effected deeply by the model linking methods,and there is indeterminacy in regional scale radiative forcing.If climate response has been considered in the modeling,there would be-10-10 W/m2 difference between two TOA radiative forcing.Surface temperature response over China area have also been compared.The following results can be figured out based on the comparison:(1) It has been clarified that there is difference in most region between results of the on-line and off-line methods both in sulfate burden and in radiative forcing at TOA and in surface temperature response.The differences on each regional averaged results are more obvious than that of large scale averaged results,which is clear and cannot be neglected.(2) T he percentage deviation of sulfate burden between the two methods is smaler than that of radiative forcing at TOA,which is smaller than that of surface temperature response.The results show indet erminacy in radiative forcing of anthropogenic sulfate and climate responses when different model linking methods have been used.The modeling results depend on the model linking methods deeply at regional scale level even,which should be emphasized in the future research work.