African swine fever remains one of the most economically important and dangerous diseases of the Suidae family. Until now, neither a safe vaccine nor a treatment against ASF has been available, which is why prevention of the disease involves biosecurity measures and early recognition based on accurate diagnosis. Nowadays, different strategies for ASF detection are discussed to reduce both animal suffering and the costs of ASF surveillance. This article aims to indicate the risk, with regard to non-invasive sampling, for the detection of ASFV. In this study, we analyzed data from three independent animal trials, in the framework of the detection of positive samples in different matrices (blood, sera, oral and rectal swabs) collected from nineteen domestic pigs infected with similar doses but under different scenarios, including different ASFV strains or routes of infection. Genetic material of ASFV was found in all matrices, but detection occurred earlier in the blood samples than in the oral and the rectal swabs. Furthermore, analyses revealed that at relevant sampling timepoints, PCR-positive blood samples were detected more frequently and reached higher percentages (up to 100% during fever) than oral and rectal swabs. Moreover, mean Ct values in blood samples collected from animals infected with virulent strains were significantly lower than in oral and rectal swabs, ensuring a higher probability of ASFV detection. High Ct values and occasional shedding in all tested matrices, in the cases of animals infected by an attenuated ASFV-strain, showed that blood sampling may be necessary to confirm the presence of anti-ASFV antibodies in sera. This study showed that during veterinary surveillance, blood sampling (for both PCR and serological analyses) is essential for the accurate diagnosis of ASF and provides the highest probability of detection of the disease.