This paper presents the investigation and classification of various 316 stainless steel grades based on their mechanical properties. The results facilitate in espousing a certain grade of steel for marine applications. The different grades have their unique chemical composition. With the advancement in soft computing, Artificial Neural Network (ANN) is used to classify the different grades of SS316 based on the changes in their mechanical properties with respect to their chemical composition and to subsequently identify the suitable grade. The twelve-different sets of chemical composition of the various grades of stainless steel were given as input to the neural network. The mechanical properties such as Ultimate Tensile strength, Yield strength and Hardness were chosen as the target. This neural network could predict the changes in the above mentioned mechanical properties due to the changes in overall chemical composition. The neural networks were trained using Back propagation Levenberg-Marquardt algorithm and it yielded best grade stainless steels as an optimal choice to be welded in Cold Metal Transfer (CMT) process. A Knowledge on these properties is inevitable to understand its behaviour for various dynamic loading conditions when subjected to joining and machining. Based on the obtained results, validations are carried out. The outcome signifies the observance of close coherence between existing data and estimated data. Considering the vitality, the same neural network is adopted and customised for Cold Metal Transfer joinability application. © IAEME Publication.