The general approach and methodology of object-oriented identification of noisy signals have been characterized. The foundation of this methodology is the concept of a noisy signal as a result of the interaction in space and time of a large number of elementary stochastic disturbances of various physical natures. Hence, the importance of solving the central boundary problem for mathematical modeling of noisy signals becomes evident. Based on the results of theoretical and experimental research, it can be asserted that the probability distributions of noisy signals are infinitely divisible. A noisy signal is regarded as an information resource of the research object. Characteristic examples of noisy signals and their corresponding identification problems in nature, technology, and biomedical technical systems have been analyzed. For instance, examples include signals in ultrasonic defectoscopy, acoustic impedance defectoscopy, vibration noise signals, electrophysiological signals generated by the electrical activity of the human brain and muscles, and more. The structural-logical scheme of object-oriented identification of noisy signals has been discussed. In this scheme, physical phenomena and technical objects define the model of the noisy signal and the corresponding set of its information features necessary for identification.