We want to test if self-similar magneto-hydrodynamic (MHD) accretion-ejection models can explain the observational results for accretion disk winds in BHBs. In our models, the density at the base of the outflow, from the accretion disk, is not a free parameter, but is determined by solving the full set of dynamical MHD equations without neglecting any physical term. Different MHD solutions were generated for different values of (a) the disk aspect ratio ($varepsilon$) and (b) the ejection efficiency ($p$). We generated two kinds of MHD solutions depending on the absence (cold solution) or presence (warm solution) of heating at the disk surface. The cold MHD solutions are found to be inadequate to account for winds due to their low ejection efficiency. The warm solutions can have sufficiently high values of $p (gtrsim 0.1)$ which is required to explain the observed physical quantities in the wind. The heating (required at the disk surface for the warm solutions) could be due to the illumination which would be more efficient in the Soft state. We found that in the Hard state a range of ionisation parameter is thermodynamically unstable, which makes it impossible to have any wind at all, in the Hard state. Our results would suggest that a thermo-magnetic process is required to explain winds in BHBs.