This paper addresses the influence of microtopography (root mean square height HRMS and correlation length LC) on RADARSAT-1 and phased array L-band synthetic aperture radar (PALSAR) backscattering coefficient (σ0) values from distinct rock alteration products of the Cu-rich district of Curaçá Valley, northeastern Brazil. The area is characterized by a semiarid environment, flat topography with rock outcrops and residual soils, and low to moderate Caatinga vegetation cover. The lithologies consist of Archean gneisses and granulites interbedded with mafic-ultramafic intrusives and upper Proterozoic marbles, schists, and phyllites. The images were acquired under distinct look azimuth and incidence angles and corresponded to four RADARSAT-1 images (F2, S2, and S7 ascending and S7 descending) and one PALSAR image (fine beam dual (FBD) descending). The research was based on the use of linear regression analyses, which showed a weak to moderate linear correlation between σ0 and HRMS and LC for both SAR data. HRMS was the most important microtopographic parameter influencing σ0, whereas LC played a secondary role. Regarding RADARSAT-1, the highest regression coefficient (R2) values were obtained for shallower incidence angles (S7), and this dependence increased from steeper to shallower incidence, regardless of changes in the look azimuth. For PALSAR, R2 was slightly higher than that for RADARSAT-1 and was related to cross-polarization. The investigation showed that backscattering for synthetic aperture radar (SAR) data from both RADARSAT-1 and PALSAR is not modulated in a predominant manner by the microtopographic variations of the geological surfaces. Pages: 262-269