We propose cooperative geographic routing (cGeo-routing) for wireless mesh networks by combining cooperative transmission with traditional geographic routing. We model and evaluate two cGeo-routing schemes including Cooperative-Random Progress Forwarding (C-RPF) and Cooperative-Nearest with Forward Progress (C-NFP). We show that cGeo-routing significantly increases the average transport capacity for a single hop in well connected mesh networks, and the gain increases with the transmitted signal-to noise ratio (SNR). Moreover, there exists an optimal topology knowledge range in C-RPF, whereas an optimal node density in C-NFP. Our results also suggest that hop-by-hop cooperation can increase transport capacity in highconnectivity and high-SNR regimes, however, it does not change the transport capacity scaling law of the mesh network.