Lunar observations by the Suomi-NPP instrument VIIRS are scheduled on a nearly monthly basis at a phase angle of approximately -51 degrees. The lunar images acquired during scheduled observations have been used for radiometric calibration stability monitoring of the reflective solar bands, band-to-band registration characterization, modulation transfer function derivation and electric crosstalk examination. A satellite roll maneuver is usually necessary for the Moon to be viewed by VIIRS detectors, which results in the loss of approximately 20-minute science data during the period. Without any scheduling, the Moon has also been regularly observed when it intrudes the field of view of the instrument’s space view port. Since the launch of Suomi-NPP in late 2011, nearly 200 unscheduled lunar observations have been made with complete lunar images captured by at least two spectral bands. These observations are made at a larger phase angle from -45 to -90 degrees and libration angle range than the scheduled lunar observation. In this paper, the strategies and methodologies of lunar calibration developed for scheduled lunar observations are applied to these unscheduled lunar observations, with necessary adaptation to account for the differences in data format. The result from the unscheduled lunar observations are provided, with the focus of it comparison with the results from scheduled lunar observations as well as solar diffuser (SD) calibration. Overall, the long-term trends of these results agree with each other and the trends from the un-scheduled lunar calibration show more fluctuation. For radiometric calibration, the difference between the lunar calibration and SD calibration strongly depends on phase angles and libration angles. If the VIIRS measurement is accurate, this indicates that the lunar irradiance reference for the radiometric calibration, modeled by the USGS robotic lunar observatory (ROLO), carries systematic error that changes with these photometric factors. An empirical correction is applied to derive the relationship between the error and the phase angle to compensate the impact. The trends after the correction shows much less fluctuation to a level similar to the trends from scheduled calibration..