The present study aims to identify the parameters from Consortium for Small-scale Modelling in CLimate Mode (COSMO-CLM) regional climate model that strongly control the prediction of extreme events, in particular, heat waves, extreme rainfall and cyclonic storms over West Bengal and the adjoining areas observed between 2013 to 2018. Metrics, namely Performance Score, Performance Index and Skill Score are employed to identify the parameters that most strongly influence the model output variables out of the 25 chosen tunable parameters corresponding to six parameterization schemes of the COSMO-CLM model. The sensitivity metrics are evaluated for three meteorological variables such as 2m-temperature, precipitation and cloud cover, simulated by the model corresponding to the different parameters for eleven extreme events over simulation domain and four inner sub-domains. It is evident from the results that only a subset of model parameters exhibits significant changes in model behaviour for distinct parameter values. In this particular study and region, no parameter is found to be sensitive from the soil parameterization scheme. Furthermore, in almost all input model parameters, the model performance reveals opposite character in different domains. Performance Index calculated from observations and model simulations with the default parameter confirms that model performance worsens when domain size reduces.