Abstract: Hot dip galvanizing of steel strip is carried out in Continuous Galvanizing Lines (CGL) in order to achieve consistency in thickness & quality of Zn-coating. The process is a complex one and is always in dynamic equilibrium depending on various process parameters. Around 20 process parameters, grouped into major sections like steel strip characteristics, pre-treatment, galvanizing process conditions, bath conditions, post treatment operations, are involved in the process. Zn-coating thickness is a nonlinear and coupled function of most of the above process input parameters. In the present work Artificial Neural Network (ANN) has been used to develop an off-line model to predict the coating thickness. Six factors that were identified through sensitivity analysis based on Taguchi’s Orthogonal Array Technique were used as independent variables in the ANN model. A Hot Dip Process Simulator (HDPS) was used to generate the input-output data with close control of input process parameters. The data set generated for training purpose has the coating thickness variation between 12 to 32 μm. The model, developed with total data-sets (234), has the prediction error of 10%. However, the maximum data-sets (183) lie in the range of 20-30 μm and model developed by taking these data-sets in this range has the prediction error of 7%. In both the cases a 6-(9-6-3)-1 network structure was used & 3/4 of data-sets were used for training & testing & 1/4 for validation.