African Swine Fever (ASF) is one of the most complex and significant diseases from a sanitary-economic perspective currently affecting the world's swine-farming industry. ASF has been endemic in Sardinia (Italy) since 1978, and several control and eradication programmes have met with limited success. In this traditional ASF endemic area, there are three susceptible host populations for this virus sharing the same habitat: wild boar, farmed domestic pigs and non-registered free-ranging pigs (known as "brado" animals). The main goal of this study was to determine and predict fine-scale spatial interactions of this multi-host system in relation to the epidemiology of ASF in the main endemic area of Sardinia, Montes-Orgosolo. To this end, simultaneous monitoring of GPS-GSM collared wild boar and free-ranging pigs sightings were performed to predict interaction indexes through latent selection difference functions with environmental, human and farming factors. Regarding epidemiological assessment, the spatial inter-specific interaction indexes obtained here were used to correlate ASF notifications in wild boar and domestic pig farms. Daily movement patterns, home ranges (between 120.7 and 2,622.8 ha) and resource selection of wild boar were obtained for the first time on the island. Overall, our prediction model showed the highest spatial interactions between wild boar and free-ranging pigs in areas close to pig farms. A spatially explicit model was obtained to map inter-specific interaction over the complete ASF-endemic area of the island. Our approach to monitoring interaction indexes may help explain the occurrence of ASF notifications in wild boar and domestic pigs on a fine-spatial scale. These results support the recent and effective eradication measures taken in Sardinia. In addition, this methodology could be extrapolated to apply in the current epidemiological scenarios of ASF in Eurasia, where exist multi-host systems involving free-ranging pigs and wild boar.