This paper presents lightning mapping array (LMA) observations of corona discharges from a windmill and its lightning protection tower in winter thunderstorms in Japan. Corona discharges from the windmill, called windmill coronas, and those from the tower, called tower coronas, are distinctly different. Windmill coronas occur with periodic bursts, generally radiate larger power, and possibly develop to higher altitudes than tower coronas do. A strong negative electric field is necessary for the frequent production of tower coronas but is not apparently related with windmill coronas. These differences are due to the periodic rotation of the windmill and the moving blades which can escape space charges produced by corona discharges and sustain a large local electric field. The production period of windmill coronas is related with the rotation period of the windmill. Surprisingly, for one rotation of the windmill, only two out of the three blades produce detectable discharges and source powers of discharges from these two blades are different. The reason for this phenomenon is still unclear. For tower coronas, the source rate can get very high only when there is a strong negative electric field, and the source power can get very high only when the source rate is very low. The relationship between corona discharges and lightning flashes is investigated. There is no direct evidence that corona discharges can increase the chance of upward leader initiation, but nearby lightning flashes can increase the source rate of corona discharges right after the flashes. The peak of the source height distribution of corona discharges is about 100 m higher than the top of the windmill and the top of the tower. Possible reasons for this result are discussed.