In general, the predictions from the use of MHV1 mixing rule (average absolute error in pressure and in vapor phase composition are 2.5% and 4.9%, respectively) is more accurate than that from WS mixing rule (6.3%, 7.9%). It is found that the less accurate results from the WS mixing rule are owing to the inherent assumptions made. By means of analyzing excess energy, we found that the accuracy from the WS mixing rule can be improved with either of the two modifications: (1) The Stavermann –Guggenheim combinatorial term in the COSMO-SAC is ignored, and so the LM is denoted as COSMOSACres. The average error in both pressure and vapor phase composition from this approach is lowered to 4.0% and 4.7%, respectively. (2) Instead of matching Gex from EOS and LM in the infinite pressure limit, we suggest to use the ambient condition instead. A detailed discussion for the improvements will be presented.