ERBB2 amplification is one of the most important and mature targets for HER2-targeted drug therapy. Somatic mutations of ERBB2 in the tyrosine kinase domain have been studied extensively, and play a role in response to anti-HER2 therapy among different cancer types. However, ERBB2 fusion has not been got attention and its relevance to HER2-targeted therapy is unclear. We comprehensively characterized ERBB2 fusions from next-generation sequencing (NGS) data between May 2018 and October 2021 in 32,131 various solid tumors. Among the tumors, 0.28% harbored ERBB2 fusions, which occurred more commonly in gastroesophageal junction cancer (3.12%; 3/96), breast cancer (1.89%; 8/422), urothelial carcinoma (1.72%; 1/58), and gastric cancer (1.60%; 23/1,437). Our population presented with a median age of 65 years (range 28 to 88 years), a high proportion of men (55 men vs 34 women; 61.80%). Among the patients with ERBB2 fusions, TP53 (82%), APC (18%), and CDK4 (15%) were the top3 co-mutant genes. What’s more, most patients with ERBB2 fusion also had ERBB2 amplification (75.28%; 67/89), which was similar to the data in the TCGA database (88.00%; 44/50). Furthermore, TCGA database shows that patients with ERBB2 fusions in pan-cancer had a worse prognosis than those without ERBB2 fusions, as well as in breast cancer. Besides, ERBB2 amplification combined with ERBB2 fusion had worse prognosis than those with only ERBB2 amplification. ERBB2 fusion may interfere the effect of anti-HER2-targeted antibody drugs and influence the prognosis of patients with ERBB2 amplification. Prospective clinical trials are warranted to confirm the results in the future.