Human brucellosis is highly detrimental, with the potential to harm various organ systems and result in long-term health issues. Brucellosis is treatable in its early stages, but becomes challenging to cure in later stages and can be fatal. Thus, early diagnosis is essential for managing the disease and minimizing complications. Traditional brucellosis detection techniques suffer from being slow and inaccurate, highlighting the need for a quicker and more reliable method. This study aimed to investigate the potential of utilizing dry serum Fourier transform infrared (FTIR) spectroscopy (absorbance/attenuated total reflectance (ATR) spectrum) in conjunction with machine learning algorithms to effectively differentiate between individuals with brucellosis and healthy control subjects. In the 1700-1500 cm-1 range, the two spectra exhibit peaks that were opposite. Among the six spectral regions examined, all regions exhibits a notably robust discriminatory capability when analyzed using support vector machine (SVM)-linear, SVM-radial basis function (RBF), principal component analysis-linear discriminant analysis (PCA-LDA), decision trees (DT), k-nearest neighbors (KNN) methodologies, and Partial Least Squares Discriminant Analysis (PLS-DA). The PLS-DA algorithm, when used with the 3080-2800 cm-1 region, achieved the highest classification accuracy. The serum FTIR (absorbance), paired with this algorithm, reaches 99.62 ± 0.52 %, surpassing the 98.48 ± 1.72 % accuracy of the ATR spectrum. According to this study, combining serum FTIR spectroscopy with the PLS-DA algorithm shows promising clinical applications for detecting human brucellosis.