Thermodynamic quantities are derived for superconducting and pseudogap regimes by taking into account both amplitude and phase fluctuations of the pairing field. In the normal (pseudogap) state of the underdoped cuprates, two domains have to be distinguished: near the superconducting region, phase correlations are important up to the temperature T φ . Above T φ , the pseudogap region is only determined by amplitudes, and phases are uncorrelated. Our calculations show excellent quantitative agreement with specific heat and magnetic susceptibility experiments on cuprates. We find that the mean field temperature T0 has a similar doping dependence as the pseudogap temperature T * , whereas the pseudogap energy scale is given by the average amplitude above Tc.One of the most intriguing problems in high temperature superconductivity is the presence of a region above the critical temperature T c and below a temperature T * where observable quantities deviate from Fermi liquid behaviour. This region is called pseudogap region [1, 2] because it contains effects similar to superconductivity like a partial suppression of electronic density of states.The origin of such a pseudogap above T c is unclear. There are four major approaches concerning its theoretical understanding: the first is based on the formation of incoherent Cooper pairs above T c . Phase order [3] or Bose condensation [4] would then establish superconductivity at T c . The second assumes that the pseudogap is induced by anti-ferromagnetic fluctuations [5]. The third approach is based on spin-charge separation where spins bind together to form spin-singlets and the energy needed to split them apart leads to the formation of a "spin-gap" [6]. The fourth assumes the existence of a quantum critical point [7] but the latter has never been observed. However, these approaches seem to be unable to describe specific heat and magnetic susceptibility.The main aim of this article is to show that various experimental observations can indeed be interpreted in terms of fluctuations of the pairing field ψ = |ψ|e iφ , and that two temperature regions have to be distinguished (see Fig. 5): for a relatively small temperature interval T c < T < T φ the phase of ψ is still correlated in space over some correlation length ξ (the Kosterlitz-Thouless correlation length in 2D). Thus, in this regime, observables are governed by correlated phase fluctuations described by the XY-model. For T φ < T < T * , phases of ψ are essentially uncorrelated (ξ is on the order of the lattice constant), but |ψ| is still fluctuating and non-zero, signaling local pair fluctuations. This explains the wide hump seen in specific heat experiments [1], the depression of the spin susceptibility [8] and the persistence of the pseudogap for T < T * .Our method has a major difference with respect to the Emery and Kivelson phase fluctuations scenario [3] of the pseudogap regime: we show that phase fluctuations influence the pseudogap only up to a temperature T φ which is much smaller than T * . Above T φ , obs...