This study presents a systematic and integrated analysis of the sensitivity of the available satellite observations to in situ soil moisture measurements. Although none of these satellites is optimized for land surface characterization, before the launches of the SMOS‐ and HYDROS‐dedicated missions they are the only potential sources of global soil moisture measurements. The satellite observations include passive microwave emissivities, active microwave scatterometer data, and infrared estimates of the diurnal amplitude of the surface skin temperature. The Global Soil Moisture Data Bank provides in situ soil moisture measurements in five separate regions. This simultaneous analysis of various satellite observations and the large amount of in situ measurements has two major advantages. First, this analysis helps identify and separate the physical mechanisms that affect the satellite observations. For example, we show that the passive microwave polarization differences at 19 GHz and above are essentially sensitive to the vegetation and not to the soil moisture (i.e., the correlation between microwave observations and soil moisture is only indirect and comes from the statistical correlation between vegetation and soil moisture). Second, this analysis enables an objective comparison of the relative potential of the various satellite observations for soil moisture retrieval when other conditions are held constant. The second part of this study benefits from this synthesis to derive a relationship between satellite observations and soil moisture at a global scale.