Major difficulties remain in the tracking of low-level clouds over land due to their size and lifetime moreoften below the spatial and temporal resolutions of current imagers, and from the underlying surface features. Cloud motion winds (CMWs) have been determined in the African monsoon flow during the summer months of 1998. Several conditions have been achieved to improve the retrieval of monsoon winds by the tracking of low-level clouds : the VIS imagery, a grid with closed grid-points (40 km), the maximal duration of efficient daylight (8h), and ten-day periods from 1st June 1998 to 30th September 1998. This retrieval has been done over areas characterized by the scarcity in time and space of conventional measurements and shows the interest of satellite measurements in such remote areas. So we used a cloud classification to validate the level of low CMWs retrieved by the Euclidean distance method as resulting from the detection of low-level clouds in the classification. The southwesterly flow has been observed penetrating into the westwards propagating cells on their southern side or turning around these cells. It can become southerly between two successive large cells. Monsoon CMWs are more numerous when the ITCZ is at its northernmost position than before or after in connection with the extension of the area covered by the rear part of the monsoon flow. The mean direction of CMWs has been measured around 220° and the mean speed around 5 ms-1 as for climatological data. Some differences are observed during the advance and the withdrawal of the monsoon flow : lower speeds with variable directions during the advance, higher speeds and more uniform direction during the retreat. Mesoscale features have been observed thanks to the high density of wind fields. Wind fields are found to be very different in two neighbouring valleys on both sides the plateau of Jos. In the Benue valley oriented south-west/north-east, monsoon CMWs remain south-westerly but can speed up to about 8 ms-1 whereas less numerous CMWs become southerly and travel at about 5 ms-1 or less in a part of the Niger valley oriented north-south. Two similar areas at about 7°N, the first one at about 2°E and the second at about 8°E, exhibit very different monsoon wind fields. Several reasons are suggested : a different upwelling in the Guinean gulf, a different convection (maybe a descending Walker branch), a different orientation of the coast line.