The Simbol-X mission, currently undergoing a joint CNES-ASI phase A, is essentially a classical X-ray telescope having an exceptional large focal length obtained by formation flying technics. One satellite houses the Wolter I optics to focus, for the first time in space, X-rays above ~10 keV, onto the focal plane in the second satellite. This leads to improved angular resolution and sensitivity which are two orders of magnitude better than those obtained so far with non-focusing techniques. Tailored to the 12 arcmin field of view and ~15 arcsec angular resolution of the optics, the ~8x8 cm2 detection area of the spectro-imager has ~ 500x500 μm2 pixels, and covers the full energy range of Simbol-X, from ~0.5 to ~80 keV, with a good energy resolution at both low and high energy. Its design leads to a very low residual background in order to reach the required sensitivity. The focal plane ensemble is made of two superposed spectro-imaging detectors: a DEPFET-SDD active pixel sensor on top of an array of pixelated Cd(Zn)Te crystals, surrounded by an appropriate combination of active and passive shielding. Besides the overall concept and structure of the focal plane including the anti-coincidence and shielding, this paper also emphasizes the promising results obtained with the active pixel sensors and the Cd(Zn)Te crystals combined with their custom IDeF-X ASICs.