In orbit satellites, the detection of anomaly is the extremely significant for ensuring the safe spacecraft; the difficult spatial temporal correlation and anomalies sparsity in data have the important challenges. In this research, developed sparse-one class-support vector machine (S-OC-SVM) method detects the anomalies in the orbit satellites. The data is collected from the sensors in the real-time environments. Then, the features are selected by using expert knowledge and determined the sequence length in preprocessing phase. Then, the features are represented by using the sparse representation method, and the anomaly detection is performed by using the proposed S-OC-SVM method. The proposed method effectively detects the anomalies in real-time with detection accuracy and precision. The S-OC-SVM technique is determined with metrics of accuracy, f1-score, precision, and recall. The S-OC-SVM technique reached 99.81% accuracy, 99.72% precision, 99.69% recall, and 99.70% f1-score that is effective than existing algorithms like improved K-nearest neighbor (KNN).