A study on the design, framework, experimental arrangement, data acquisition, and subsequent analysis of cold crack detection using fiber Bragg grating (FBG) sensors during welding is presented. Cold crack induced acoustic emission (AE) events are detected by multiple FBG sensor configurations and using a commercial FBG AE interrogation system. The crack locations are theoretically predicted using a method of linear localization based on the differences in the time of arrival of the AE events at different FBG sensors. The approach is experimentally validated by employing packaged and unpackaged FBG sensor arrays in a welding experiment with gas tungsten arc welding on ultrahigh toughness high hardness armor steel. Frequency analysis of the AE events is also carried out to understand the cold crack characteristics. The weld region of the steel plate is diced into 1-mm cross-sectional pieces, and the positions of the cold cracks are identified using scanning electron microscope imaging and then cross correlated with the measured AE events, and thereby confirming the correct identification of cold crack location using the FBG sensing method.