In Morocco, Agriculture is a key sector of the national economy, playing crucial social and economic roles. The sector accounts for around 14 to 20% of the Gross Domestic Product (GDP) and represents 43% of all employment. Winter cereals (soft wheat, durum wheat and barley) are produced all over the country, occupying nearly 65% of agricultural lands and therefore cereal yields forecasting is a major tool for decision making, allowing for planning in advance actions like annual cereal imports or aids to farmers. The present study highlights the substantial contribution of remote sensing (RS) and Geographic Information Systems (GIS) techniques in predicting soft wheat yields at the rural commune of Ouled Saleh, Region Casablanca-Settat in Morocco. The forecasting methodology was based on two steps: First, a land cover map of the study area was produced using Sentinel imagery to identify agricultural zones; second, simple linear regression models were established between the Normalized Difference Vegetation Index (NDVI), derived from the Moderate Resolution Imaging Spectrometer (MODIS) and soft wheat yields over the period 2002-2012. Our results showed high correlations between the NDVI of agricultural lands, averaged over the period from February till March or April and soft wheat yields. Therefore, NDVI can be used as a predictor index to early forecast soft wheat yields one to two months before harvest.