Human Immunodeficiency Virus (HIV) belongs to the Lentivirus genus, Retroviridae family, enveloped by a lipid bilayer within which the capsid protein encases the viral genome, reverse transcriptase, and integrase proteins, key components for viral replication. Viral capsid has been linked to key early and late stages of viral infection, including nuclear entry, promoting reverse transcription and assembly of new viral particles within target TCD4+ lymphocytes. Effective treatments for HIV involve multi drug therapy, which can reduce the patient's viral load to undetectable values, thus avoiding the appearance of Acquired Immunodeficiency Syndrome (AIDS). In this study, a conserved region of the HIV capsid protein was selected and 84 compounds were selected from a massive Artificial Intelligence-based virtual screening as potential HIV capsid assembly disruptors. In vitro screening was performed using recombinant protein and complemental approaches were carried out to identify molecules capable of interfering with capsid multimerization. From this work, 9 compounds were selected as successful to continue through in cell and toxicity assays for further development as possible HIV treatments. In conclusion, this work demonstrates the efficiency of integrating rational computational and experimental methodologies to identify new candidates as potential antiviral molecules.