Abstract: Objective To evaluate the clinical value of half-Fourier acquisition single-shot turbo spin sequence based on deep learning reconstruction algorithm (HASTE-DL) T2WI sequence in MRI of pancreas. Methods Sixty-nine patients underwent pancreatic conventional BLADE-TSE-T2WI and based on deep learning HASTE-DL-T2WI sequences scanning using 3.0T MRI. The overall quality, pancreatic sharpness, biliary duct clarity and artifacts of the two groups were qualitatively scored by the Likert 5-point scale method for subjective evaluation. The contrast to noise ratio (CNR) and signal to noise ratio (SNR) of normal tissue and the lesion of the pancreatic of the two groups of images were measured and analyzed. The scanning time of the two sequences were also recorded. Results The overall image quality of the HASTE-DL sequence, the scores of pancreatic sharpness and biliary clarity were better than those of BLADE-TSE sequence (P<0.001).There was no statistically significant difference in the scores of image artifacts between the two groups (P >0.05). The SNR of normal pancreatic tissue, the SNR of lesion and CNR of HASTE DL sequence were better than those of BLADE-TSE sequence (P<0.001). The scanning time of HASTE-DL sequence was 78% shorter than that of BLADE-TSE sequence. Conclusion Compared with BLADE-TSE sequence, the overall quality of HASTE-DL is better, and the pancreatic sharpness and bile duct display clarity are higher, SNR and CNR are better, and the scanning time is shorter. Therefore, HASTE-DL T2WI sequence has good clinical application value in the MRI of pancreas.