Hypertension is a leading cause of cardiovascular disease and premature death worldwide and it puts a heavy burden on the healthcare system. It is, therefore, very important to detect and evaluate hypertension and related cardiovascular events as to for efficient diagnosis, treatment and management. Hypertension can be evaluated with noninvasive cardiac signals, such as electrocardiogram (ECG) and photoplethysmogram (PPG) signals. Most of the previous studies predicted hypertension from ECG and PPG signals with extracted features that are correlated with hypertension. However, correlation is sometimes unreliable and may be affected by confounding factors. In this study, we propose a causal inference based approach to identify feature variables from ECG and PPG signals that are potentially causally related with hypertension. The method of greedy equivalence search was employed to construct the causal graph of features and hypertension. With causal features identified from the causal graph, we used machine learning models to diagnose hypertension. The machine learning classification models achieve great classification performance, among which random forest model has the best classification performance, with accuracy being 0.987, precision being 0.990, recall being 0.981, and F1-score being 0.985. The results show that the causal inference based approach can effectively predict hyper-tension.Clinical relevance- This paper proposes a new hypertension risk prediction method, which uses causality instead of correlation as the feature screening criteria to establish a causal graph of hypertension, which can predict the hypertension more reliably.