Male fertility is significantly influenced by the quality of the semen. It has gradually declined in the past few years, and changes in lifestyle. Numerous studies have shown that poor lifestyle choices are a major contributing factor to male reproductive diseases and low-quality semen. Machine Learning (ML) is well-suited to handle the dynamic interactions that exist between predictor traits and outcomes across large datasets. However, selecting the most important features from huge datasets becomes a very difficult task. In this paper, Butterfly Optimization Algorithm (BOA) is introduced to find the environmental factors and lifestyle choices that impact seminal quality. The collective behaviours of foraging and mate-finding in butterflies served as the model for BOA. The BOA, which computed the average training loss reduction due to feature utilisation for each dataset, was used to determine the feature significance. BOA is utilised to find more pertinent features has the impact on the seminal quality. Input features were considered categorical features, and the output features were considered dichotomous features according to the Feed-Forward Neural Network (FFNN) classifier. Dataset is collected from University of California Irvine (UCI). The assessment measures include precision, sensitivity/recall; specificity, f-measure, and accuracy have guided the experimentation analysis. FFNN is compared to other methods like Clustering Based Decision Forest (CBDF), K-Nearest Neighbors (KNN), and Multi-Layer Perceptron (MLP).