The success of in vitro fertilization (IVF) treatments is influenced by a complex interplay of multiple factors, including patient-specific biochemical parameters. In this study, a machine learning approach is used to analyze the biochemical markers of the patient undergoing IVF. A dataset of 28 patients undergoing IVF treatments was collected, comprising a range of 21 biochemical parameters. Traditional logistic regression, support vector machines, decision trees and random forests classification were applied to analyze and model the data. The feature selection and dimensionality reduction techniques has been used to identify the most relevant and informative markers for IVF prediction. Subsequently, the various sets of selected marker values have been used to train and predict IVF success outcomes, and to evaluate performance of classification models in terms of accuracy, sensitivity, and specificity depending on the selected markers or features. The results have shown that it is possible to formulate a certain probability of IVF based on these markers, and that most of the used classification models required a smaller number of markers. As for the performance, the best results were achieved by the SVM and decision trees approaches, which achieved 70-80 % prediction accuracy using several parameters less than 6.