A proposed model of the human respiratory system is described and analyzed in this article. We introduce a simple measurement technique which can track transient changes in human respiratory system. The present work provides a methodology for auto regressive exogenous (ARX) modelling of human respiratory system. The proper model structure and model parameters are determine for the human respiratory system. Estimated   model parameters will reflect the dynamic changes in respiratory system so the differences between normal and disease person can be detected using this model. The results are promising and models obtained for physiological system are able to describe the difference between the normal patient and disease patent by the diagnosis of change in the model parameters.