Partially cloud-filled pixels can be a significant problem for remote sensing of cloud properties. For example, optical depths and effective particle sizes are often too small or too large, respectively, when derived from radiances that are assumed to be overcast but contain radiation from both clear and cloud areas within the satellite imager field of view. Few methods have been developed for taking these effects into account, yet the results from such retrievals can bias the statistics of cloud properties making it difficult to match observations with model calculations of cloud parameters. One possible means for reducing the impact of such partially cloud filled pixels is to estimate the cloud fraction within each pixel. Although the nominal resolution for most channels on the Geostationary Operational Environmental Satellite (GOES) imager and the Moderate Resolution Imaging Spectroradiometer (MODIS) on Terra are 4 and 1 km, respectively, both instruments also take visible (VIS, 0.65 µm) channel data at 1 km and 0.25 km, respectively. Thus, it may be possible to obtain an improved estimate of cloud fraction within the lower resolution pixels by using the information contained in the higher resolution VIS data. The GOES and MODIS multisopectral (0.64, 3.7, 11, and 12 µm ) data are analyzed with the algorithm used for the Atmospheric Radiation Measurement Program and the Clouds and Earth’s Radiant Energy System to derive cloud amount, temperature, height, phase, effective particle size, optical depth, and water path. Normally, the algorithm assumes that each pixel is either entirely clear or cloudy. In this study, a threshold method is applied to the higher resolution VIS data to estimate the partial cloud fraction within each low-resolution pixel. The cloud properties are then derived from the observed low-resolution radiances using the cloud cover estimate to properly extract the radiances due only to the cloudy part of the scene. This approach is applied to both GOES and MODIS data to estimate the improvement in the retrievals for each resolution. The 1-km MODIS data are also used to simulate 4-km GOES pixels to determine how well the multi-resolution approach applied to 4-km data compares to analyses of actual 1-km data. This technique is most likely to yield improvements for low and midlevel layer clouds that have little thermal variability in cloud height.