Without any effective tool for screening and early diagnosis, lung cancer shows the highest mortality in cancer related death. Here, we show the development of lung cancer proteome biomarkers and in vitro diagnostics based on multivariate index assay (IVD-MIA). By LC-ms/ms–based proteomics approaches combined with glycoprotein enrichment, low molecular weight protein enrichment technologies in the sera of the cancer patients, and secretome analysis from primary cultured lung cancer and normal tissues, we have discovered various potential lung cancer protein biomarkers. TMT or iTRAQ-based quantitative and label-free proteomics combined with fucosylated glycoprotein enrichment approaches also revealed that not only the amount of the glycoprotein biomarkers but also their fucosylation levels and patterns can serve as diagnostic and prognostic serological markers for lung cancers. Fucosylated protein biomarkers were validated by lectin-hybrid ELISA and immunoassays. We also validated the biomarkers by multiple reaction monitoring (MRM) in the sera of the patients. Functional analysis also revealed that biomarker PON1 promotes ROS deregulation protecting the mitochondria from dysregulation. SAA and QSOX1 promote lung cancer metastasis by immunomodulating macrophages. A logistic regression model based on SID-MRM assay of SERPINA4 and PON1 serum protein markers with age factor revealed AUC 0.92 for differential diagnosis between lung cancer and other lung diseases. We have developed various pairs of biomarker-specific monoclonal antibodies and used these antibodies which can be used for lateral flow assay and microfluidics assay. Using three selected potential protein biomarkers and age, a deep learning algorithm showed AUC 0.90 in the sera test of 1,700 lung cancer patients and normal control. Our future development of IVD-MIA will further improve the detection and diagnosis of lung cancers.