This study aimed to identify potential biomarkers and the therapeutic targets for colorectal adenocarcinoma by systematically evaluate a large scale of long noncoding RNAs (lncRNAs) expression data from TCGA. The algorithm t-distributed stochastic neighbor embedding and hierarchical clustering were utilized to group the samples into three clusters that showed a different prognosis. To identify the relationship between the clustered groups and different histoclinical features, different statistical methods were used. The functions of LINC01234 and MIR210HG were investigated with the help of the public database. The results showed that the expression levels of lncRNAs were able to distinguish the tumor samples from the normal tissues and in further they were able to predict the prognosis of the patients. We proposed two potential lncRNAs, which might serve as a biomarker or therapeutic targets. LINC01234 can be a good biomarker. In contrast, MIR210HG participated in the progression of colorectal adenocarcinoma by regulating hypoxia. It might function through an lncRNA-microRNA-messenger RNA regulatory network with MIR210 and RASSF7.