Haloxylon ammodendron and Haloxylon persicum are important drought-tolerant plants in northwest China. The whole-genome sequencing of H. ammodendron and H. persicum grown in their natural environment is incomplete, and their transcriptional regulatory network in response to drought environment remains unclear. To reveal the transcriptional responses of H. ammodendron and H. persicum to an arid environment, we performed single-molecule real-time (SMRT) and Illumina RNA sequencing. In total, 20,246,576 and 908,053 subreads and 435,938 and 210,334 circular consensus sequencing reads were identified by SMRT sequencing of H. ammodendron and H. persicum, and 15238 and 10135 unigenes, respectively, were successfully collected. In addition, 9794 and 7330 simple sequence repeats (SSRs) and 838 and 71 long noncoding RNAs were identified. In an arid environment, the growth of H. ammodendron was restricted; plant height decreased significantly; and the base and branch diameters became thinner. Hydrogen peroxide (H ₂O ₂) content and peroxidase activity were increased. Under dry and wet conditions, 11,803 and 15,217 differentially expressed genes (DEGs) were identified in H. ammodendron and H. persicum, respectively. There were 319 and 415 DEGs in the signal transduction pathways related to drought stress signal perception and transmission, including the Ca ²⁺ signal pathway, the ABA signal pathway, and the MAPK signal cascade. In addition, 217 transcription factors (TFs) and 398 TFs of H. ammodendron and H. persicum were differentially expressed, including FAR1, MYB, and AP2/ERF. Bioinformatic analysis showed that under drought stress, the expression patterns of genes related to active oxygen (reactive oxygen species) scavenging, functional proteins, lignin biosynthesis, and glucose metabolism pathways were altered. This is the first full-length transcriptome report concerning the responses of H. ammodendron and H. persicum to drought stress. The results provide a foundation for further study of the adaptation to drought stress. The full-length transcriptome can be used in genetic engineering research.