I briefly review the cosmological constant problem and the issue of dark energy (or quintessence). Within the framework of quantum field theory, the vacuum expectation value of the energy momentum tensor formally diverges as k⁴. A cutoff at the Planck or electroweak scale leads to a cosmological constant which is, respectively, 10¹²³ or 10⁵⁵ times larger than the observed value, Λ/8πG 10⁻⁴⁷ GeV⁴. The absence of a fundamental symmetry which could set the value of Λ to either zero or a very small value leads to the cosmological constant problem. Most cosmological scenarios favour a large time-dependent Λ-term in the past (in order to generate inflation at z 10¹⁰), and a small Λ-term today, to account for the current acceleration of the universe at z 1. Constraints arising from cosmological nucleosynthesis, CMB and structure formation constrain Λ to be sub-dominant during most of the intermediate epoch 10¹⁰ < z < 1. This leads to the cosmic coincidence conundrum which suggests that the acceleration of the universe is a recent phenomenon and that we live during a special epoch when the density in Λ and in matter are almost equal. Time varying models of dark energy can, to a certain extent, ameliorate the fine-tuning problem (faced by Λ), but do not resolve the puzzle of cosmic coincidence. I briefly review tracker models of dark energy, as well as more recent brane inspired ideas and the issue of horizons in an accelerating universe. Model independent methods which reconstruct the cosmic equation of state from supernova observations are also assessed. Finally, a new diagnostic of dark energy—statefinder— is discussed.