Satellite-based monitoring of crop phenology is commonly built on the analysis of Vegetation Index (VI) time series by extracting phenological metrics. The relatively fine detection of the various timings in crops growth during their development cycle depends on the density and regularity of valid observations. Medium spatial resolution (MSR) daily observations provide consistent cloud-free n-day composite time series, suitable for phenological applications, but do not offer an adequate spatial resolution. MSR pixels are generally mixed pixels or "mixels", composed of several land cover classes, which complicate crop-specific monitoring from space. To address the MSR mixel problem, we implemented a spatial disaggregation (SD) approach that estimates a cropspecific VI based on the crop fraction in a mixel provided by a land use map. First, SD was applied on synthetic MSR data (i.e. Sentinel-2 data aggregated at 300 m) in order to test the method in an ideal case. After validation, the method was applied to PROBA-V data, using 300 m and 10-day composites over a large area around Paris, for four main crops (i.e. winter cereals, spring barley, oilseed rape and maize) in 2016-2017. The evaluation of SD was done by comparing disaggregated data with reference data (i.e. Sentinel-2 10 m). Indeed, two main results were observed, i) SD was able to reconstruct the crop-specific VI time series of all crops and ii) PROBA-V data increased the number of crop-specific VI valid observations at certain stages of the crop's growth period compared to Sentinel-2 data, this with a consistent and regular revisit throughout the growth cycle. In conclusion, SD can be used to improve the exploitation of MSR data in seasonal crop monitoring, especially during the transition periods when the VI of crops are likely to change quickly. This paves the way for monitoring crop phenology over fragmented landscapes, from sensors such as MODIS or SPOT-VEGETATION, even for years before Sentinel-2 launch.