This paper presents a rule-based machine learning model for financial fraud detection without using data resampling techniques. The model is designed to identify patterns in financial transactions using a set of decision rules, making it more interpretable and explainable than other machine learning models. The study compares the proposed rule-based model with several existing machine learning models, including random forest (RF), decision tree (DT), multi-layer perceptron (MLP), k-nearest neighbor (KNN), naive Bayes (NB), and logistic regression (LR), using two benchmark datasets: PaySim and BankSim. The results show that the proposed rule-based model outperforms these models in terms of accuracy, precision, recall, and other evaluation metrics, achieving accuracy and precision of 0.99 on both datasets. The model generates association rules based on transaction patterns and evaluates them using support and confidence measures. The rules are then refined and validated to ensure their effectiveness in detecting fraudulent transactions. The study highlights the potential of rule-based models in financial fraud detection, particularly in handling imbalanced datasets without the need for resampling. The proposed model demonstrates high performance, with a Matthews correlation coefficient (MCC) of 0.993 for the PaySim dataset and 0.995 for the BankSim dataset. The model's ability to generate and validate rules effectively makes it a promising approach for financial fraud detection.This paper presents a rule-based machine learning model for financial fraud detection without using data resampling techniques. The model is designed to identify patterns in financial transactions using a set of decision rules, making it more interpretable and explainable than other machine learning models. The study compares the proposed rule-based model with several existing machine learning models, including random forest (RF), decision tree (DT), multi-layer perceptron (MLP), k-nearest neighbor (KNN), naive Bayes (NB), and logistic regression (LR), using two benchmark datasets: PaySim and BankSim. The results show that the proposed rule-based model outperforms these models in terms of accuracy, precision, recall, and other evaluation metrics, achieving accuracy and precision of 0.99 on both datasets. The model generates association rules based on transaction patterns and evaluates them using support and confidence measures. The rules are then refined and validated to ensure their effectiveness in detecting fraudulent transactions. The study highlights the potential of rule-based models in financial fraud detection, particularly in handling imbalanced datasets without the need for resampling. The proposed model demonstrates high performance, with a Matthews correlation coefficient (MCC) of 0.993 for the PaySim dataset and 0.995 for the BankSim dataset. The model's ability to generate and validate rules effectively makes it a promising approach for financial fraud detection.