Anti-programmed cell death-1 (PD-1)/PD-ligand-1 (PD-L1) treatments are effective in a fraction of patients with advanced malignancies. However, the majority of patients do not respond to it. Resistance to cancer immunotherapy can be mediated by additional immune checkpoints. We hypothesized that co-targeting of PD-L1 and lymphocyte-activation gene 3 (LAG-3) could provide an alternative therapeutic approach. Here, we developed IBI323, a dual blockade bispecific antibody targeting PD-L1 and LAG-3. We assessed the binding affinity, blocking activity, cell bridging effect, and immunomodulation function of IBI323 using in vitro assays. We also evaluated, in two humanized mouse models, anti-tumor effects and antitumor T cell immunity induced by IBI323. IBI323 bound to PD-L1 and LAG-3 with similar potency as its parental antibodies and blocked the interaction of PD-1/PD-L1, CD80/PD-L1, and LAG-3/MHC-II. Moreover, IBI323 mediated the bridging of PD-L1+ cells and LAG-3+ cells and demonstrated superior immune stimulatory activity compared to each parent antibody in mixed leukocyte reaction. In PD-L1/LAG-3 double knock-in mice bearing human PD-L1 knock-in MC38 tumors, IBI323 showed stronger anti-tumor activity compared to each parental antibody. The better antitumor response correlated with increased tumor-specific CD8+ and CD4+ T cells. IBI323 also induced stronger anti-tumor effect against established A375 tumors compared with combination in mice reconstituted with human immune cells. Collectively, these data demonstrated that IBI323 preserved the blockade activities of parental antibodies while processing a novel cell bridging function. Based on the encouraging preclinical results, IBI323 has significant value in further clinical development.