Early diagnosis of hepatic cancer is a major public health challenge. While changes in serum N-glycans have been observed as patients progress from liver fibrosis/cirrhosis to hepatocellular carcinoma (HCC), the predictive performance of N-glycans is yet to be determined for HCC early diagnosis as well as differential diagnosis from liver fibrosis/cirrhosis. In a total sample of 247 patients with hepatitis B virus-related liver disease, we characterized and compared the serum N-glycans in very early/early and intermediate/advanced stages of HCC and those with liver fibrosis/cirrhosis. Additionally, we performed a retrospective timeline analysis of the serum N-glycans 6 and 12 months before a diagnosis of the very early/early stage of HCC (EHCC). A predictive model was built, named hereafter as Glycomics-EHCC, incorporating the glycan peaks (GPs) 1, 2, and 4. The model showed a larger area under the receiver operating characteristic curve compared with a traditional model with the alpha-fetoprotein (0.936 vs. 0.731, respectively). The Glycomics-EHCC model had a sensitivity of 84.6% and specificity of 85.0% at a cutoff value of -0.39 to distinguish EHCC from liver fibrosis/cirrhosis. Moreover, the Glycomics-EHCC model was able to forecast a future EHCC diagnosis with a sensitivity and specificity over 90% and 85%, respectively, using the serum N-glycan biosignatures 6 or 12 months earlier when the patients were suffering from liver fibrosis/cirrhosis before being diagnosed with EHCC. This serum glycomic biosignature model warrants further clinical studies in independent population samples and offers promise to forecast EHCC and its differential diagnosis from liver fibrosis/cirrhosis.