Methamphetamine (METH) is a commonly abused psychostimulant with a high addictive nature. The paraventricular nucleus of thalamus (PVT), a key nucleus for arousal, has attracted much attention in the reward process of substance use. However, at which stage dose the PVT encode the reward process? How to reduce the side-effects of modulating PVT on wakefulness during the treatment of substance use? These issues remain unclear. The goal of the current study is to explore the role of the PVT and the glutamatergic projections from medial prefrontal cortex (mPFC) to PVT in the reward process of METH. Here, the conditioned place preference (CPP) was used to assess the reward process of METH in male mice, combined with methods of c-Fos mapping, virus-based neural tracing, patch-clamp recording, EEG-EMG recordings, optogenetics and designer receptor exclusively activated by designer drugs (). The glutamatergic neurons in PVT (PVT) were triggered during METH CPP-Test, rather than by METH CPP-Training. Suppressing either PVT or glutamatergic projection from mPFC to PVT efficiently disrupted the acquisition of METH CPP in male mice, mainly mediated by the GluN2A subunit of NMDA receptor. Further, inhibition of PVT affected the rhythm of EEG-EMG, whereas inhibition of glutamatergic projection from mPFC to PVT did not. PVT is involved in the reward process of METH at the retrieval stage of METH-conditioned context, rather than at the stage of encoding association between METH and context. The glutamatergic projections from mPFC to PVT, especially the GluN2A molecule, may be a promising therapeutic target for reducing METH reward, as there are no significant side effects on wakefulness.