Background: Immune checkpoint inhibitors (ICIs) have improved survival outcomes and resulted in long-term responses in primary liver cancer in some patients. However, its efficacy is limited by the risk of tumor recurrence, resulting in rapid death and graft loss if patients are not selected appropriately. Therefore, it is necessary to identify patients suitable for such therapy. Methods: 215 patients with primary liver cancer with immunotherapy were screened between August 2018 and October 2020 as a training set and our validation set were included 71 patients of hepatocellular carcinoma from Jiangxi Cancer Hospital, The Second Affiliated Hospital of Nanchang Medical College from May 2019 to July 2021. The primary endpoint was the disease control rate (DCR), and the secondary endpoints were overall survival (OS) and progression-free survival (PFS). Results: In the traing set, Neutrophil-leukocyte ratio (NLR) >3, Alpha-fetoprotein (AFP) level >20 ng/ml, distant metastasis at baseline, and absence of combination with targeted therapy were associated with non-DCR in the training set. Moreover, NLR >3 and AFP level >20 ng/ml were also independently associated with OS. Furthermore, a hepatic immune predictive index (HIPI) based on NLR >3 and AFP level >20 ng/mL was developed and associated with worse clinical outcomes. In validation set, HIPI was associated with overall survival. Conclusion: Baseline hepatic immune predictive index based on NLR and AFP level is an effective indicator in ICI-treated patients with primary liver cancer. Our findings may help guide the selection and on-treatment strategies for immunotherapies for primary liver cancer patients.