BackgroundThe geographical distribution of HIV in Hainan Province is highly heterogeneous. Yet the patterns of cross-regional transmission remain unclear. To address this issue, we integrated HIV transmission network analysis with spatial analysis to characterize cross-regional transmission patterns across the province.MethodsWe analyzed 2,135 HIV-1 pol gene sequences obtained from HIV/AIDS patients reported in Hainan Province between 2021 and 2025. A molecular transmission network was constructed using a genetic distance threshold of 1.5% to identify cross-regional links among network nodes. Spatial autocorrelation analysis was subsequently performed to characterize the transmission patterns within and across cities in Hainan Province.ResultsAmong 2,135 HIV-1 pol sequences, 1,396 (65.39%) sequences were included in the molecular transmission network. Older aged (≥50 years), no drug resistance, and infection with CRF07_BC and CRF65_cpx were independently factors associated with a higher likelihood of clustering. Global spatial autocorrelation analysis indicated significant spatial dependence in clustering rates across Hainan (Moran’s I = 0.256, P < 0.05). Local analysis identified Wuzhishan as a high-low cluster for all subtypes, and Danzhou specifically for CRF07_BC. Network linkage analysis revealed intensive intracity transmission within Haikou and Sanya across all subtypes, with the strongest intercity links between these two cities for CRF07_BC, CRF01_AE, and CRF65_cpx. Of the 1,396 nodes, 79.87% displayed intercity linkage, and female sex was less likely linked, while infection with CRF65_cpx increased linkage likelihood.ConclusionThis study characterizes the patterns of cross-regional HIV transmission in Hainan Province, highlighting the Haikou-Sanya link as a crucial intervention point. Enhanced regional surveillance and coordinated prevention strategies are recommended for this transmission route.