The purpose of this study was to compare and evaluate the performance of electronic leaf wetness duration (LWD) sensors in measuring LWD in a cotton crop canopy when unpainted and painted sensors were used. LWD was measured with flat, printed-circuit wetness sensors, and the data were divided into two periods of 24 days: from 18 December 2001 to 10 January 2002, when the sensors were unpainted, and from 20 January to 13 February 2002, when the sensors were painted with white latex paint (two coats of paint). The data analysis included evaluating the coefficient of variation (CV%) among the six sensors for each day, and the relationship between the measured LWD (mean for the six sensors) and the number of hours with dew point depression under 2 °C, used as an indicator of dew presence. The results showed that the painting markedly reduced the CV% values. For the unpainted sensors the CV% was on average 67% against 9% for painted sensors. For the days without rainfall this reduction was greater. Comparing the sensor measurements to another estimator of LWD, in this case the number of hours with dew point depression under 2 °C, it was also observed that painting improved not only the precision of the sensors but also their sensitivity, because it increases the ability of the sensor to detect and measure the wetness promoted by small water droplets.