Large clusters, similarly to solids, couple very efficiently to intense subpicosecond laser pulses. Near 100 % of the laser radiation can be absorbed leading to the observation of highly charged ions with energies reaching MeV and electrons with energies up to a few keV. One fascinating feature of this interaction is its efficiency for converting photons in the eV range to x-rays with keV energies. The study of x-ray emission allows to investigate this interaction on a very short time scale. Indeed, even though only ions with inner-shell vacancies contribute to X-ray spectroscopy, their observation gives access to the dynamical evolution of the irradiated cluster on a time scale comparable to that of the laser pulse duration. As an example, excited states of argon ions, from Ar12+ up to Ar16+, with a K-shell vacancy are produced in laser heated clusters : those states have lifetimes that can be as low as a few tens of femtoseconds (15 fs for the Ar16+ 1P1 state). Since the inner-shell vacancies are produced by electron-impact ionisation, x-ray spectroscopy can provide insight into the electron dynamics and more precisely on the heating mechanisms, which allow electrons to gain energy as high as the inner-shell binding energies. KeV X-rays have been studied from rare-gas clusters irradiated with intense (I <1017 W cm-2) infrared and blue laser pulses of 50 – 2000 fs duration. (Ar)n, (Kr)n and (X^e)n clusters with n > 104 (i.e. 10-40 nm of diameter) have been investigated. Quantitative measurements have been performed on the evolution of absolute photon yields and charge state distributions of the emitting ions with different physical parameters governing the interaction ; namely intensity, polarisation, pulse duration and wavelength of the laser as well as the size of the clusters. Among the most significant results obtained so far, we show evidence of low laser intensity threshold values for keV X-ray production, an optimum heating time when mapping the cluster dynamics by varying the pulse duration at constant laser energy and a saturation in the X-ray emission probability above a given cluster size. The production of inner-shell vacancies being very sensitive to energetic electrons produced during the interaction, the local cold plasma approximation used in the well known “nanoplasma” model is no longer valid. Our findings initiated theoretical developments using microscopic calculations based on Monte Carlo simulations. This new theoretical treatment leads to a very good agreement with the experimental data on the evolution of the X-ray yields as the function of the laser intensity. Analysis of the evolution of the high-energy part of the kinetic energy distribution of the heated electrons as a function of pulse duration and cluster size is in progress. All the experiments have been performed at the LUCA (Laser Ultra Court Accordable) facility in Saclay- France (CEA).