Autophagy may provide the source of nutrients for tumor cells. We aim to develop an autophagy-related signature to predict the progression from lower-grade gliomas (LGG) to glioblastoma multiforme (GBM) and prognosis. Totally, 686 differentially expressed genes (DEGs) and 73 long non-coding RNAs (DELs) were identified between GBM and LGG samples from the Chinese Glioma Genome Atlas (CGGA). Of them, 131 DEGs were intersected with autophagy genes from the Human Autophagy Database; while 54 DELs co-expressed with autophagy-related DEGs. Ten autophagy-related genes were associated with overall survival and could distinguish GBM from LGG, with the accuracy of 0.891 using CGGA dataset and 0.790 using The Cancer Genome Atlas (TCGA) dataset. The risk score was established based on these 10 genes. Patients with higher risk score were at an increased risk of developing GBM (49.7% vs. 21.3%; p < 0.001) and worse prognosis than those in low risk group. The prognostic accuracy was 0.840 and 0.744 for CGGA and TCGA dataset, respectively. Age, recurrence, isocitrate dehydrogenase mutation and risk score were independent prognostic factors and thus they were used to build a nomogram which showed the highest prognostic power. This established nomogram may aid the clinical decision making of personalized treatment..