This study aimed to elucidate the potential genes of the matrix metalloproteinase (MMP) family, responsible for the progression of laryngeal squamous cell carcinoma (LSCC). Besides, we ascertained the changes in common malignant behaviors in vitro by knocking down MMP1. TCGA, GEO, Oncomine, and microarray data were conducted to analyze the expression levels of MMPs and to find tissue-specific genes in LSCC. Univariate and multivariate Cox regression analyses were established in the construction of a prognostic model based on expression profiles and clinical information of LSCC in TCGA. We then comprehensively analyzed survival, co-expression network, and immune infiltration based on a prognostic model by Kaplan-Meier analysis, WGCNA, and CIBERSORT. Thereafter, qRT-PCR, proliferation, Transwell, and wound-healing assays were used to assess the accuracy of the bioinformatics data. A total of seven genes in the MMP family were identified as differentially expressed genes (DEGs) by integrating three public databases and microarray data. Additionally, multivariate Cox regression was used to establish a four-gene (MMP1/3/8/10) prognostic model, which exhibited a better predictive accuracy than the TNM (tumors/nodes/metastases) based model. The prognostic model was related to plasma cells, CD8+ T cells, follicular helper T cells, resting NK cells, and M0 macrophages infiltration. The expression of MMP1, MMP3, and MMP10 was the highest in head and neck squamous cell carcinoma (HNSC) compared to other cancer in the Oncomine and GEPIA dataset. Further, MMP1 demonstrated significant upregulation in 40 paired LSCC tissues. Eventually, MMP1 downregulation inhibited cell viability, colony formation, and cell migration in TU686 and FaDu cells. Our findings suggest that the four-gene signature might be associated with the prognosis. Further, we revealed that MMP1 is a pivotal biomarker for the biotherapy and prognostic evaluation of patients with LSCC.