In this study, an online monitoring technique for continuous fatigue crack quantification and remaining fatigue life estimation is developed for plate-like structures using nonlinear ultrasonic modulation and artificial neural network (ANN). First, multiple aluminum plates with different thicknesses were subjected to cyclic loading tests with a constant amplitude, and ultrasonic responses were obtained from three PZT transducers placed on each specimen. Second, an ANN is constructed by (1) defining the specimen thickness, the elapsed fatigue cycles, and two features extracted from the ultrasonic responses, named as cumulative increase and decrease of nonlinear modulation components, as inputs and (2) the crack length and the remaining fatigue life as outputs. The results of validation tests indicate that the proposed technique can estimated the crack length and the remaining fatigue life with a maximum error of 1.5 mm and 2 k cycles, respectively. The uniqueness of this technique lies on (1) fatigue crack quantification and remaining fatigue life estimation using nonlinear ultrasonic modulation, and (2) data-driven continuous crack quantification and prognosis.