In this study, HYBRID and RMSD error functions are used to calculate the error deviation between the experimental and predicted data for carbon based composite/CO2 and AQSOA-ZO1/water pairs. For fitting, six isotherm models, namely Dubinin-Astakhov (D-A), Modified D-A, Tòth, Langmuir, Freundlich, and Hill model are selected for the Type-I isotherm, whilst Mahle, Modified Langmuir, Sun and Chakraborty, Guggenheim-Anderson-de Boer (GAB), Hybrid model (Henry+Sips) and D-A models are selected for the Type-V isotherm. Based on error values, Tòth model provides less error compared to other models for Type-I, likewise Sun and Chakraborty model for Type-V isotherm. Seven information criterion (IC's) are employed to predict the optimum isotherm model for the studied pairs. Here, 1000 bootstrap samples are taken to measure the accuracy of sample estimation and error is measured for each sample. Hence, the error distribution is found with average error and confidence interval (CI). Using average error, all IC's are calculated. On the basis of IC's, Tòth model provides less information loss compared to other models for Type-I isotherm, similarly Sun and Chakraborty model provides less information loss for Type-V isotherm. This analysis found that using bootstrap HYBRID mean error Sun and Chakraborty model reduces 4% more information loss compare to modified Langmuir model of Type-V isotherm. Introduction In order to standardized experimental data of different adsorbent/refrigerant pairs, it is important to correlate with different isotherm models. Generally, isotherm data are correlated applying various isotherm models where the best fitted model is used to analyse the system performance. For instance, Pal et al. [1] used Tόth and modified D-A models for CO2 adsorption onto composite, Kayal et al. [2] used D-A, modified Langmuir, and Sun & Chakraborty models for water adsorption onto AQSOA zeolite. Muttakin et al. [3] fitted different isotherm models for different types of IUPAC classified adsorption isotherms. Kim et al. [4] fitted Hybrid model, which was a combination of Henry and Sip's models, for water adsorption onto ferroaluminophosphate. But, still there is a lack of information about the selection of the most suitable model for different types of isotherm also the reason behind selection in terms of statistical point of view. In this study, statistical information criteria using bootstrap sample has been used to select the optimized isotherm model for Type-I and Type-V isotherms. Two sets of equilibrium adsorption data, CO2 onto a carbon based consolidated composite [1] and water onto AQSOA zeolite [2], are investigated to find the optimum isotherm model. Two error functions, RMSD and HYBRID, are used to optimize the fitting parameters. Using R-programming, bootstrap samples of N = 1000 are taken and model selection criteria are used for finding optimum isotherm model.