Abstract: To address the issues of high cost and strong subjectivity associated with manual observation in dairy cow estrus detection, as well as the high computational power consumption caused by high-dimensional model features, this study proposes a dairy cow estrus detection method based on unsupervised clustering and interpretable feature selection. The method utilizes the K-means algorithm to cluster raw activity time series data collected by neck-mounted accelerometers on dairy cows to automatically generate estrus behavior labels. Three models are compared: K-Nearest Neighbors (KNN), Support Vector Machine (SVM), and Random Forest (RF). These models are evaluated under time windows of 10 , 30 s, and 50 s. Finally, the optimal model is determined by the Friedman test and the Nemenyi post-hoc test. The Shapley Additive Explanations (SHAP) method is introduced to quantify feature contributions and screen for the key feature. The results indicate that the RF model based on a 10 s time window achieves the best performance. Through SHAP selection, the feature is reduced from 25 to 3. The performance of the dimension-reduced model shows no significant deviation, with accuracy, precision, recall, and F1 score reaching 0.9911, 0.9370, 0.9955, and 0.9654, respectively. In independent trials based on three-axis acceleration data from dairy cows, the RF model achieves an estrus detection accuracy of 0.9090 using a 10 s window following dimensionality reduction. The proposed method enables both estrus label generation and feature dimensionality reduction, while maintaining model interpretability. This provides technical support for developing low-power, long-endurance estrus monitoring devices for dairy cows.