In the era of radio astronomy, the high sensitivity of the Square Kilometre Array (SKA) could play a decisive role in the detection of new radio sources. In this work, we study the SKA sensitivity to the synchrotron radio emission expected by the annihilation of TeV DM candidate in the Draco dwarf spheroidal galaxy. On the one hand, we consider model-independent DM candidates: we find out that with 1000 hours of data-taking, SKA1-MID will be able to exclude up to 10 TeV thermal DM candidates that annihilate in $W^+W^-$ and $bbar b$ channels. We also study as these constraints improve by including a density enhancement due to a DM-spike associated with an intermediate-mass black hole in Draco. On the other hand, we consider extra-dimensional brane-world DM candidates, dubbed branons. In this specific scenario, SKA allows us to set constraints on the branon parameter space ($f$,$M$), where $f$ is related to the coupling of the branon to the Standard Model particles and $M$ is the mass of the branon itself. In particular, we consider two different branon DM candidates. We find out that SKA will be able to set more stringent constraints on the branon DM candidate required in order to fit the AMS-02 data, yet the sensitivity of the instrument should be improved in order to study the branon candidate for the Galactic Centre. Nonetheless, we show that SKA represents - among other detectors - the most promising instrument for multi-wavelength detection of synchrotron radio emission by annihilating multi-TeV DM.