Objective The aim of this study was to optimize the cut-off value of phenylalanine (Phe) for phenylketonuria (PKU) screening in Xinjiang Uygur Autonomous Region based on the time of blood sampling. Study Design In this study, 110,806 neonates born in 91 obstetrics and gynecology hospitals of Xinjiang Uygur Autonomous Region between June 2017 and December 2019 were divided into two groups (i.e., groups 1 and 2) based on the sampling time. The concentration of Phe was determined using fluorimetric method. The optimization of the Phe cut-off value was conducted using the receiver operating characteristic curve from the treating set involving 80,354 neonates. Then, the diagnostic values of the optimized Phe cut-off value were evaluated using validation set involving 30,452 neonates, based on the comparison of sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) obtained from conventional cut-off value. Results A range of cut-off values was used for preliminary Phe concentrations in the two groups to analyze the sensitivity, specificity, PPV, and NPV. The optimized cut-off value of Phe in group 1 was 2.0, while that in the group 2 was 2.21. A comparison was given to PPV, NPV, sensitivity, and specificity generated by the optimized cut-off value and the conventional cut-off value, which yielded similar sensitivity, specificity, and PPV, and less recalled number of samples. Conclusion The optimization of cut-off value of Phe based on sampling time is feasible for PKU screening in Xinjiang Uygur Autonomous Region. In addition, the false positive rate was significantly reduced, which may save more efforts in sample recalling process. Key Points