In the network virtualization envi-ronments, one of the most challenges is how to map the virtual networks (VNs) onto a shared substrate network managed by an infrastruc-ture provider (InP), which is termed as virtual network embedding problem. Prior studies on this issue only emphasize on maximizing the revenue or minimizing the energy consump-tion while ignoring the reliability requirements of end-users. In our work, we incorporate the reliability probability into the virtual network embedding process with an aim to improve the QoS/QoE of end users from a new perspective. We devised two novel reliable virtual network embedding algorithms called RRW-MaxMatch and RDCC-VNE based on RW-MaxMatch and DCC-VNE, respectively. Extensive sim-ulations demonstrated that the efficiency of our proposed algorithms is better than those of two primitive algorithms in terms of the reli-ability demands, the acceptance ratio of virtual networks and the long-term average revenue.