Phase change memory (PCM) is an emerging non-volatile data storage technology concerned by the semiconductor industry. To improve the performances, previous efforts have mainly focused on partially replacing or doping elements in the flagship Ge-Sb-Te (GST) alloy based on experimental "trialand-error" methods. Here, the current largest scale PCM materials searching is reported, starting with 124 515 candidate materials, using a rational highthroughput screening strategy consisting of criteria related to PCM characteristics. In the results, there are 158 candidates screened for PCM materials, of which ≈68% are not employed. By further analyses, including cohesive energy, bond angle analyses, and Born effective charge, there are 52 materials with properties similar to the GST system, including Ge 2 Bi 2 Te 5 , GeAs 4 Te 7 , GeAs 2 Te 4 , so on and other candidates that have not been reported, such as TlBiTe 2 , TlSbTe 2 , CdPb 3 Se 4 , etc. Compared with GST, materials with close cohesive energy include AgBiTe 2 , TlSbTe 2 , As 2 Te 3 , TlBiTe 2 , etc., indicating possible low power consumption. Through further melt-quenching molecular dynamic calculation and structural/electronic analyses, Ge 2 Bi 2 Te 5 , CdPb 3 Se 4 , MnBi 2 Te 4 , and TlBiTe 2 are found suitable for optical/electrical PCM applications, which further verifies the effectiveness of this strategy. The present study will accelerate the exploration and development of advanced PCM materials for current and future big-data applications.