Our aim is to relate classical cardiovascular signal analysis to models of the cardiovascular (CV) and control systems taking into account its multiple feedback loop organisation. We discuss knowledges on the short-term control of CV system, focusing on the analysis of CV time-series and control of arterial pressure and RR-interval. In the first and second sections, we introduce modelling concepts which lead us to the definition of a possible decentralised control of CV system, with several baroreflex control loops, and at a time-discrete approach for the estimation of their sensitivities. We consider two types of vascular compartment: a passive one composed of elastin and collagen fibers, and an active one characterised in addition by smooth muscle. Passive vascular compartment modelling leads us to introduce the classical Windkessel model, widely applied for blood pressure curve analyses. We demonstrate that Windkessel model can be derived from a distributed model of the vascular compartment by a three-point discretisation of the one-dimensional partial differential equations of blood flow. We relate the definition of the vascular resistance and compliance to physical peculiarities of the local vessel wall. Looking for a possible mechano-chemical command of the heart function, we conceive a hierarchical control of the heart, separating the mechanical from the chemical control of the cardiac pump. We model the aortic and carotid baroreceptors, considering the three different steps of transduction: baroreceptor sensitivity to the inside pressure, pressure-deformation transduction process and deformation neural firing rate transduction process. The model is based on accumulated physiological and histological evidences and incorporates a local sympathetic feedback mechanism. We demonstrate that the knowledge of the history of the time derivative of the vessel wall stress is rich enough to allow the estimation of the inner baroreceptor pressure without the need of a pressure set-point. Active vascular compartment modelling leads us to consider blood pressure in the compartment as an output of the model. Whereas state variables are blood volume of the vascular compartment, the strain and stress of smooth muscles which are controlled by the autonomic nervous system (ANS) and by a local control of blood oxygenation. Finally, we conceive the plan of a possible CV controller. In our point of view, the ANS regulates blood flow and pressure in different parts of the CV system with local feedback loops supposed to be fast and efficient. We demonstrate the necessary existence of negative and positive feedbacks: negative for blood pressure and/or flow, positive for blood flow. In the third section, we introduce actual knowledges and clinical relevance of (i) the origin of CV variability and (ii) of baroreflex mechanism, relating these concepts to CV modelling. In the last sections, we develop classical methods of CV spectral and time varying analysis, applying them in clinical conditions. A joint time-frequency distribution, the Smoothed Pseudo Wigner-Ville Distribution (SPWVD) and an harmonic analysis method, the Complex Demodulation (CDM), are evaluated on synthetic and real CV time-series. We show the ability of this SPWVD-CDM to estimate the instantaneous amplitude and frequency of CV oscillations and the instantaneous phase between respiratory and CV oscillations.