In contrast with most static studies in nuclear medicine, dynamic studies have an inherently high noise level because of low count statistics. Preprocessing of the raw image data by the application of suitably designed smoothing filters, provides a significant contribution to the dynamic studies. Because dynamic studies are 3-D, consisting of two spatial dimensions and time, preprocessing of the raw image data by a specially tailored 3-D filter, would be expected to fit more precisely and efficiently than the utilization of conventional 2-D filters. In this study, a new 3-D Double Window Adaptive Hybrid Smoothing Algorithm (3-D DWAH filter) is introduced and the results obtained by its application on the pixel basis dynamic renal scintigraphic images are compared with those of the conventional filters and unsmoothed raw data. The 3-D DWAH filter effectively and optimally combines the FIR-Median Hybrid (FMH) and the running mean filters within a variably tuned interval providing a significant improvement in SNR without unacceptable degradation of spatial and temporal resolution.