In the thesis, we consider the blind maximum-likelihood detection problem for distributed space-time block coded cooperative communication systems. The randomized space-time coding strategy proposed by Mergen & Scalione [8] is applied in the relay network, and the efficient semidefinite relaxation (SDR) based blind ML detector [13] is used in the destination receiver. Applying the blind detection techniques to the environment of distributed cooperative communications not only overcomes the fast fading channel effects but also provides robustness against the effective channel variations due to the changing in network topology (i.e., any termination of the collaborated relays). In the system, the source node transmits the information data to the destination receiver via a group of relays. By using the so-called nonintersecting subspace orthogonal space-time block codes (NIS-OSTBCs), the unique data identifiability can be guaranteed under the blind detection scenario. Besides, we proved that the blind ML detector obtain the maximum transmit diversity gain at the destination receiver. The final simulation results also demonstrate that the proposed system yields promising bit error performance. Extension of the proposed blind ML receiver to other distributed space-time coding strate- gies are also investigated in the remaining part of the thesis.