Centrifugal ultrafiltration after methanol extraction of whole plasma was used as an optimal condition for the preparation of blood plasma before metabonomic studies. The plasma samples from 102 lung cancer patients and 34 healthy volunteers were prepared with this approach. With ultra-performance liquid chromatography（UPLC） coupled with quadrupole time-of-flight mass spectrometry（Q-TOF MS） analysis, the samples were investigated in order to find potential disease biomarkers. After data acquisition, orthogonal signal correction partial least squares models were built to differentiate the healthy volunteers from lung cancer patients and to identify metabolites that showed significantly different expression between the two groups. Several metabolite ions were identified as potential biomarkers according to the variable importance in the project（VIP） value in both ion modes. Five lysophosphatidylcholines were further identified as specifically lysoPC 16：0, isomer of lysoPC 16：0, lysoPC 18：0, lysoPC 18：1 and lysoPC 18：2. These results suggest that UPLC coupled with Q-TOF MS is an effective technique for the analysis of plasma metabolites in metabonomic studies.
