Studies on ultraviolet light-emitting diode (UV-LED) water disinfection have shown advantages, such as safety, flexible design, and lower starting voltages. However, information about reactivation after UV-LED disinfection is limited, which is an important issue of UV light-based technology. In this study, the photoreactivation and dark repair of Escherichia coli after UV-LEDs and low pressure (LP) UV disinfection were compared. Four UV-LED units, 265 nm, 280 nm, the combination of 265 + 280 (50%), and 265 + 280 (75%) were tested. 265 nm LEDs was more effective than 280 nm LEDs and LP UV lamps for E. coli inactivation. No synergic effect for disinfection was observed from the combination of 265 and 280 nm LEDs. 265 nm LEDs had no different reactivation performances with that of LP UV, while 280 nm LEDs could significantly repress photoreactivation and dark repair at a low irradiation intensity of 6.9 mJ/cm2. Furthermore, the UV-induced damage of 280 nm LEDs was less repaired which was determined by endonuclease sensitive site (ESS) assay. The impaired protein activities by 280 nm LEDs might be one of the reasons that inhibited reactivation. A new reactivation rate constant, Kmax, was introduced into the logistic model to simulate the reactivation data, which showed positive relationship with the maximum survival ratio and was more reasonable to interpret the results of photoreactivation and dark repair. This study revealed the distinct roles of different UV lights in disinfection and reactivation, which is helpful for the future design of UV-LED equipment. Display omitted •265 nm LEDs was more effective for E. coli disinfection.•No synergic effect for disinfection from the combination of 265 and 280 nm LEDs.•280 nm LEDs significantly reduced the reactivation processes.•Protein-damaging may be one reason that inhibited the reactivation.•New reactivation rate constant with clear physical meaning was introduced.