An alternative method for dealing the feature selection problem is proposed through improving the MFOalgorithm by using the OBL approach to generate the population initialization and the DE approach as a methodfor the local searching for the MFO scheme. Due to that, the MFO explores the search space well better thanexploiting it; so, this combination improves the efficiency of the MFO and avoids it from getting stuck in localpoints. According to these strategies (OBL and DE), the proposed method is called OMFODE. Our method hasthree phases, initialization phase, phase of updating, and classification phase. Such phases of the proposed OMFODEapproach are given with more details in the following subsections (see Fig. 1).