We compare the rate of kinetic energy injected by stellar winds into the extragalactic H II region NGC 2363 to the luminosity needed to feed the observed temperature fluctuations. The kinetic luminosity associated to the winds is estimated by means of two different evolutionary synthesis codes, one of which takes into account the statistical fluctuations expected in the Initial Mass Function. We find that, even in the most favorable conditions considered by our model, such luminosity is much smaller than the luminosity needed to account for the observed temperature fluctuations. The assumptions underlying our study are emphasized as possible sources of uncertainty affecting our results.