Hypertension is one of the most common cardiovascular diseases threatening people's health worldwide. Although, hypertension itself is rarely an acute problem, it increases the risk of cardiovascular events and kidney diseases. Recent studies have shown that the end-organ damage associated with hypertension is more strongly correlated with ambulatory blood pressure monitoring (ABPM) than with traditional clinic BP measurements. Currently, the ambulatory, cuff-based devices are predominantly based on automatic techniques which are inherently motion-sensitive. The other devices which claim to compensate for motion artefacts measure the arterial pressure at the wrist and are less accurate than cuff-based measurement. This research aims to develop a cuff-based ABPM technique which can measure BP accurately during their daily lives. The primary objective of this research is to investigate two major issues related to ABPM technique: (1) the theory of oscillometric BP measurement method used in most of ABPM device; (2) the method to compensate for the noises during the measurement. This thesis introduces for the first time a 3D finite element (FE) model which simulates the entire oscillometric BP measurement process. The model is validated by both arm simulator and clinical results. The brachial artery closure process and the factors of arm material properties in BP measurement are discussed. The model indicates that the nonlinearity of brachial artery plays the key role in oscillometric BP measurement. It also offers a new explanation of the common phenomenon: overestimation in the elderly and underestimation in the younger. Since BP pulse, arm motion and environmental tremors all lead to the upper arm skin stretches, it is anticipated that the strain distribution due to these causes is unique for different arm motion. This thesis also describes the design of a piezoelectric strain sensor array and the relevant method for eliminating the noises and determining the subject’s BP. A FE model, whose geometry is obtained from Visible Human Body dataset, is established to study surface strain distribution during different arm motions. Referring to its results, the piezoelectric sensor array is designed and used in the clinical experiments. Using the obtained signals, a generalized input-output configuration of the designed measurement system is developed. The transfer functions of the system are determined through empirical equations. Using the developed method, the device is able to detect the arm motions, compensate for the noises and determine patients’ BP. The results can be used as a guide for developing a new type of ABPM device insensitive to artefacts.