The COVID-19 pandemic is almost half year old now and is still tormenting the humans to unimaginable extent with deeper interference to their routine life and peace. As approved vaccines are yet to be synthesized and standard therapeutic procedures are awaiting establishment for fighting against new Corona virus, several treatment modalities are being suggested and tried out by scientific community. Many of such approaches follow a drug repurposing approach as a possible remedy could prevent a great amount of loss in a shorter span of time. In this background, we report our attempt made for identifying a solution to this malady with a similar strategy. We used machine learning algorithms and the structural information of already approved drugs to identify potential therapeutics for managing the Covid-19 crisis. The experiments have been done with a group of 77 antiviral molecules (for the training phase of machine learning) and another group, comprising 9 antivirals and 11 antimalarials (meant for the testing phase). All the chosen molecules are approved category drugs and have significant drug action against the viruses. The identified molecules are subjected to validation by making docking studies with recently released crystal structures of Corona Virus. The binding affinity of the tested small molecules with three selected severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) structures are computed and compared with the affinity scores of five other medications viz. Hydroxychloroquine, Favipiravir, Dexamethasone, Dichlorobenzyl alcohol and Amyl metacresol followed by subjecting the results to the statistical test of ANOVA. The predicted therapeutics in conjunction with their already established characteristics could be further put to evaluation by approved clinical trials towards determining the efficiency of them against COVID-19 infection.