We are developing an algorithm for off-line detection and classification of sleep-disordered breathing based on time series analysis of nasal mask pressure and acoustic respiratory input impedance measured by forced oscillation technique at a frequency of 20 Hz throughout the night. A first version of the algorithm was applied to a data set consisting of full-night measurements on 5 subjects. The data set had a total duration of 34 hours and contained 577 respiratory events (hypopneas, obstructive and central apneas) recognized by the staff physicians of an accredited sleep laboratory. The algorithm detected 455 (79 %) of these events and 138 events that had not been marked by the physicians. 75 % of the congruently detected events were also concordantly classified. After further optimization and evaluation, this approach might be useful when implemented into a device designed to screening or treatment control of sleep-related breathing disorders at home.