A noninvasive technique for monitoring tissue temperature changes due to heating fields using diagnostic ultrasound is described in this paper. The approach is based on the discrete scattering model used in the tissue characterization literature and the observation that most biological tissues are semi-regular scattering lattices. It has been demonstrated by many researchers and verified by us that the spectrum of the backscattered radio frequency (RF) signal collected with a diagnostic ultrasound transducer from a semi-regular tissue sample exhibits harmonically related resonances at frequencies determined by the average spacing between scatterers along a segment of the A-line. It is shown theoretically and demonstrated experimentally (for phantom, in vitro, and in vivo media) that these resonances change with changes in the tissue temperature within the processing window. In fact, changes in the resonances (delta f) are linearly proportional to changes in the temperature (delta T), with the proportionality constant being determined by changes in the speed of sound with temperature and the linear coefficient of thermal expansion of the tissue. Autoregressive (AR) model-based methods aid in the estimation of delta f. It should be emphasized that this new technique is not a time of flight velocimetric one, so it represents a departure from previously used ultrasonic methods for tissue temperature estimation.