New virulence races of the wheat leaf rust pathogen emerge continuously, due to frequent mutations in pathogenicity, often resulting in loss of resistance in wheat cultivars originally resistant to leaf rust. So, it is important to develop a specific molecular marker for diagnosis and prediction of wheat leaf rust in the field. On the basis of our previous studies of completing transcriptome sequencing of wheat leaf rust fungus PHNT, in this study we selected candidates of effectors from the wheat leaf rust fungus PHNT transcriptome database, used real-time fluorescence quantitative PCR (qPCR) to detect their expression patterns at different time points after leaf rust fungus infection with wheat, then designed primers according to the sequences of the selected candidates of effectors based on expression differences, conducted PCR amplification using genome DNA of PHNT as templates to screen a specific molecular marker for wheat leaf rust. As a result, a total of 24 candidates of effectors were screened based on the transcriptome database, and 17 expressed significantly difference by qPCR. The PCR results showed that the candidates of effector PTTG-05290 had a specific band of 920bp appearing in the wheat leaf rust fungus PHNT, and the specific band was detected in 25 leaf rust races, but was absent in six rust fungi of wheat stripe rust, bean rust, jujube rust, willow rust, apple rust and reed rust, which indicated that the marker was specific for wheat leaf rust fungus. In addition, the results of qPCR showed that the expression level of PTTG-05290 is 218 times higher at the expression peak (4 dpi) than that at 0 dpi, and the gene could be detected in large quantities at 0.5 dpi. So, the specific molecular marker can be used for the detection and prediction of leaf rust in the field at an early stage, providing a shortcut for the adoption of wheat cultivars of resistance to leaf rust.