Long non-coding RNAs (lncRNAs) have been shown to be vital players in a majority of physiological and pathological processes, including tumorigenesis and tumor progression. The aim of this study was to identify lncRNAs that can serve as biomarkers for cancer prognosis. Based on dosage sensitivity, we utilized the biological features of known cancer-related lncRNAs, and identified microRNA and transcription factor (miRNA-TF) co-regulatory motifs in an effort to establish a holistic analysis framework and predict new cancer prognosis-associated lncRNAs. We found that lncRNAs with low dosage sensitivity regulated by more than 3 types of co-regulatory motifs were more likely to be associated with cancer. By the use of the integrative analysis of 3035 tumor samples across 9 types of cancer, a total of 33 cancer prognosis-associated lncRNAs were identified. Additionally, on the basis of the miRNA-TF co-regulatory network, we also predicted potential small molecule drugs such as Glucocorticoid and Ginsenoside Rh2 for treating KIRC by targeting miRNA. This study explains the causes of abnormalities in the genome from a new perspective, and provides new clues for cancer diagnosis and prognosis, and research for anti-cancer drugs.