Objectives: To introduce the concepts of improper m− polar soft fuzzy graphs and related terms. Methods: An algorithm has been established to apply m− polar soft fuzzy graph to make a decision for medical diagnosis in the current COVID-19 scenario. It has been demonstrated with examples. Findings: The newly integrated ideas of the soft sets and m− polar fuzzy sets will lead to numerous prospective applications in the m− polar fuzzy set theoretical domain by adding extra fuzziness in analysing. m− polar soft sets that are most useful in practical applications. The concepts such as improper m− polar soft fuzzy graphs, totally improper m − polar soft fuzzy graphs, neighbourly improper m− polar soft fuzzy graphs, neighbourly totally improper m− polar soft fuzzy graphs, highly improper m− polar soft fuzzy graphs and highly totally improper m− polar soft fuzzy graphs are defined. Novelty: The identical condition of neighbourly improper 3 -polar soft fuzzy graphs and highly improper 3-polar soft fuzzy graphs are discussed and validated. Various results related to these concepts have been established.