Acoustic emission (AE) signals generated by microcrack formation in front of a propagating fatigue crack in a thin‐plate compact tension specimen of 7075‐T6 aluminum alloy were analyzed to investigate the dynamics of microcrack formation. The dominant wave motion was in‐plane so that the dynamic elastic wave field could be analyzed by generalized plane stress theory. The acoustic signals were detected by an array of four miniature piezoelectric transducers which were mounted on the edge of a specimen ahead of the advancing crack. A second‐rank force moment tensor was used to model the microcrack formation and a time‐domain double‐iterative signal deconvolution method was used to process the detected acoustic signals to remove the effects of the structure and to recover the temporal and spatial characteristics of the AE source. It was found that a force moment tensor model can be used to represent the growth of the microcrack.