An audio data processing device and method includes selecting one multiple-S-curves, and running an interpolation arithmetic upon original points according to the selected multiple-S-curve. Selecting one multiple-S-curve includes selecting one of a plurality of S-curves corresponding to a current utilization ratio of a processor, obtaining audio data having a fixed length including sampling frequencies and the original points of the obtained audio data, and selecting the one multiple-S-curves corresponding to the selected S-curve and the obtained sampling frequencies.