Location-Based Social Network （LBSN） combines social network with geographical locations， providing users with novel personalized experiences. The protection of user location privacy is crucial for the secure operation of LBSN systems. To address the problem of rigid location privacy protection methods leading to low data utility and decreased quality of Location-Based Service （LBS） experiences， a User Preference-based and Differentially Private Location Privacy Protection （UPDP-LPP） method was proposed. Firstly， the set of user stay points was obtained by using a stay point extraction algorithm. Secondly， the types of stay points were labeled by using a feature fusion method. Finally， by dynamically obtaining privacy budget and noise sensitivity through user preferences， Laplace noise was added to the privacy radius to protect sensitive user location information. Experimental results on two public real datasets show that the proposed method improves the data utility of privacy protection by more than 10% compared to TLDP （Trajectory Location Data Protection）， DPLPA （Differential Privacy-based Location Privacy protection Algorithm）， and LPPM （Location Privacy Protection Mechanism） when the privacy budget is the same. It can be seen that UPDP-LPP not only protects user location privacy， but also enhances data utility.