Human respiratory syncytial virus (HRSV) infection is a significant cause of morbidity and mortality, particularly among the children and the elderly. Despite extensive efforts, there is currently no formally approved vaccine and effective antiviral options against HRSV infection are limited. The development of vaccines and antiviral strategies for HRSV was partly hampered by the lack of an efficient lethal mouse model to evaluate the efficacy of the candidate vaccines or antivirals. In this study, we established a lethal HRSV mouse model by consecutively passaging a clinical HRSV isolate, GZ08-0. GZ08-18 was isolated from mouse bronchioalveolar lavage fluids at the 50th passage of GZ08-0. Importantly, all GZ08-18-inoculated mice succumbed to the infection by day 7 post infection, whereas all GZ08-0-inoculated mice recovered from the infection. Subsequent investigations demonstrated that GZ08-18 replicated to a higher titer in mouse lungs, induced more prominent lung pathology, and resulted in higher expression levels of a number of key pro-inflammatory cytokines including IFN-γ, MIP-1α, and TNF-α in comparison to that of GZ08-0. The cyclophosphamide pretreatment rendered the mice more susceptible to a lethal outcome with less rounds of virus inoculation. Full genome sequencing revealed 17 mutations in GZ08-18, some of which might account for the dramatically increased pathogenicity over GZ08-0. In addition, by using ribavirin as a positive control, we demonstrated the potential application of this lethal mouse model as a tool in HRSV investigations. Overall, we have successfully established a practical lethal mouse model for HRSV with a mouse-adapted virus, which may facilitate future in vivo studies on the evaluation of candidate vaccines and drugs against HRSV.