This study aimed to determine the effects of irrigation methods, including reservoir-drainage, wick irrigation and drip irrigation, on the growth, water holding capacity of substrate and nutrient uptake of lettuce ($Lactuca$ $sativa$ L. var. $cripspa$), chicory $Cichorium$ $intybus$ L. var. $folisum$ Hegi) and endive ($Cichorium$ $endivia$ L. var. $endivia$) grown in an extensive green roof system. Two independent experiments were carried out during summer and autumn seasons with and without supplying mineral nutrients. In both experiments, the ranking of substrate water holding capacity was in the order of drip irrigation > wick irrigation > reservoir-drainage method. Only the wick irrigation method constantly maintained the water content in the substrate during the growing period. In the summer experiment, not only fresh and dry weights and relative growth rate (RGR) of all three vegetables, but also total nitrogen (T-N) and magnesium (Mg) contents of the leaves were the highest at the wick irrigation treatment. The contents of phosphorous (P) and potassium (K) in the leaves were the highest at the drip irrigation treatment. In the autumn experiment, shoot fresh weights of all three vegetables were the highest at the drip irrigation treatment, while shoot dry weights of lettuce and chicory were the highest at the wick irrigation treatment. The RGR of lettuce and chicory were the highest at the drip irrigation treatment, while the RGR of endive was the lowest at the wick irrigation treatment. In the both experiment, dry matter contents (DMs) were the highest at the reservoir-drainage treatment. There was no effect of irrigation treatments on the photosynthesis, stomatal conductance and transpiration rate of the lettuce and chicory. These results suggest that the wick irrigation is more efficient for growing leafy vegetables in the extensive green roof system compared to the reservoir-drainage and drip irrigation.