The high photostability of DNAs and RNAs is inextricably related to the photochemical and photophysical properties of their building blocks, nucleobases and nucleosides, which can dissipate the absorbed UV light energy in a harmless manner. The deactivation mechanism of the nucleosides, especially the decay pathways of cytidine (Cyd), has been a matter of intense debate. In the current study, we employ high-level electronic structure calculations combined with excited state non-adiabatic dynamic simulations to provide a clear picture of the excited state deactivation of Cyd in both gas phase and aqueous solution. In both environments, a barrierless decay path driven by the ring-puckering motion and a relaxation channel with a small energy barrier driven by the elongation motion of C = O bond are assigned to <200 fs and sub-picosecond decay time component, respectively. The presence of ribose group has a subtle effect on the dynamic behavior of Cyd in gas phase as the ribose-to-base hydrogen/proton transfer process is energetically inaccessible with a sizable energy barrier of about 1.4 eV. However, this energy barrier is significantly reduced in water, especially when an explicit water molecule is present. Therefore, we argue that the long-lived decay channel found in aqueous solution could be assigned to the Cyd-water intermolecular hydrogen/proton transfer process. The present study postulates a novel scenario toward deep understanding the intrinsic photostability of DNAs and RNAs and provides solid evidence to disclose the long history debate of cytidine excited-state decay mechanism, especially for the assignment of experimentally observed time components.