Soil moisture has always been an important surface variable, but its retrieval in a global fashion has always proved very difficult. From dedicated science networks to operational ones, it is almost impossible to access a global view of our planet, not to mention measurement homogeneity. Remote sensing can provide a solution in the sense that the coverage can be global and the uniqueness of the instruments favors homogeneity. However, the requirements are very stringent and only “all weather” direct measurements could provide an answer. Currently, this is only possible with low-frequency, passive microwave measurements. Moreover, these measurements only concern the surface layer. To access the vadose zone, indirect methods using models and assimilation are required. Finally, specific surface conditions (dense rain forest, snow covered soils, etc.) can significantly hamper the measurements, making global coverage a difficult goal to attain. Considering the importance of soil moisture in a range of fields, it seems important to see how, by merging different approaches, we can try to offer a complete coverage and provide long temporal series. The second goal is to provide a feeling of what could be the future evolution required to obtain, from space, accurate, adequate, and useful information on soil moisture.