A two-hop neighborhood information-based real-time routing design is proposed for wireless sensor networks. The approach of mapping packet deadline to a velocity is first adopted in SPEED routing. However, this routing decision is made based on the two-hop velocity. An efficient probabilistic drop is used to enhance efficiency while reducing deadline miss ratio when there is no forwarding candidate that can have the required velocity. If the deadline is not stringent, a cost function is embedded that can release the nodes frequently chosen to be the forwarder. An improvement of energy consumption balance is achieved across the network. The true characteristics of physical and MAC layers are captured in the simulation. A real lossy link model is drawn from experiments through Mica2 Motes. Simulation results show that compared with the existing SPEED-S that only utilizes one-hop information the proposed routing scheme has a lower deadline miss ratio and higher energy efficiency and it does not degrade the delay performance in general. The proposed design can be applied in real-time applications based on sensor networks which are more demanding in service quality.