<p>Hotel recommendation (HR) for tourists is an obvious multi-attribute group decision-making (MAGDM) problem, which is a very important part of tourist travel. To address the uncertainty of information in HR, probabilistic picture hesitant fuzzy set (PPHFS) has been applied as an effective mathematical tool in this study to express decision information. This article proposes a novel MAGDM technique for HR. Firstly, to address the issue of inconsistent lengths of possible positive-membership hesitant degree (PPOMHD), possible neutral-membership hesitant degree (PNEUMHD), and possible negative-membership hesitant degree (PNEGMHD) for two PPHFEs that require normalization, we propose a normalized method named probability splitting algorithm to avoid the drawback of changing the original information by adding new elements based on the risk preferences of decision-makers in the past. Then, the basic operation of the PPHF elements (PPHFEs) does not satisfy the defect of closeness, we propose some novel operations for aggregating PPHFEs. In addition, based on these operations, the PPHF weighted average (PPHFWA) and PPHF weighted geometric (PPHFWG) operators were proposed, and their excellent properties were studied in detail. To better measure the difference between two PPHFEs, we have also proposed several novel PPHF distance measures (PPHFDisMs) for PPHFEs. Considering the irrational behavior of decision-makers (DMs), we extend the behavioral TOPSIS (BTOPSIS) technique to the PPHF environment and propose the PPHFWA and PPHFWG based BTOPSIS (PPHFBTOPSIS) technique, which is applied to HR. Finally, the robustness of the proposed PPHFBTOPSIS technique is demonstrated through parameter analysis.</p>