The aim of this study is to train neural network which predict thermodynamic design data without involving role of thermodynamic properties, temperature concentration complex relationship and formula of coefficient of performance. Authors have presented values of weights and bias obtained from the trained ANN through which one can formulate mathematical equations for predicting thermodynamic design data. Authors have achieved coefficient of multiple determinations R2 as 0.99999, 0.99997 and 0.99953 for solution concentration in absorber, generator and enthalpy-based coefficient of performance respectively which are quite satisfactory to validate ANN approach for this study.Keywords: Absorption refrigeration system, ANN, thermodynamic design data, solution concentration, ammonia-sodium thiocyanate
