Alpha-fetoprotein (AFP) computational secreted network construction and analysis of human hepatocellular carcinoma (HCC) is very useful to identify novel markers and potential targets for prognosis and therapy. By integration of gene regulatory network infer and the database for annotation, visualization, and integrated discovery, we identified and constructed significant molecule AFP secreted network from 25 no-tumor hepatitis/cirrhotic liver tissues and 25 HCC patients in the same GEO Dataset GSE10140-10141. Our result verified AFP secreted module in the upstream of no-tumor hepatitis/cirrhotic liver tissues (AMELY, LCN2, and REG3A activation; DKK1, SFRP4, and SPINK1 inhibition) and its downstream (PRSS1, REG3A, and TSHB activation; AMELY and DKK1 inhibition), and also in the upstream of HCC (LCN2, REG3A, and SFRP4 activation; AMELY and DKK1 inhibition) and its downstream (AMELY activation; DKK1, LCN2, PRSS1, SEMA3B, and SPINK1 inhibition). Importantly, we data-mined that AFP secreted cluster of HCC is involved in disease mutation (only in HCC terms) without cell surface receptor linked signal transduction, neuroactive ligand-receptor interaction, cell-cell signaling, and pancreas (only in no-tumor hepatitis/cirrhotic liver tissues terms), the condition which is vital to invasion of HCC. Our result demonstrated that common terms in both no-tumor hepatitis/cirrhotic liver tissues and HCC include secreted extracellular region, extracellular region part, extracellular space, signal peptide, signal, disulfide bond, glycosylation site N-linked (GlcNAc...), and glycoprotein, and these terms are less relative to invasion; therefore, we deduced the weaker AFP secreted network in HCC consistent with our number computation. We predicted AFP high expression localization within cells of HCC and without secretion to extracellular matrix. It would be necessary of AFP secreted function to decrease invasion of HCC.