The present study was an attempt to investigate the effects of premodified input, interactionally modified input and modified output on 80 EFL learners' comprehension of new words. The subjects were randomly assigned into four groups of pre modified input, interactionally modified input, modified output and unmodified (control) groups. Each group included 20 students. The participants in the experimental and control groups were taught under different conditions. In each session a matrix picture of a place and its furniture and equipments was given to the participants and they were trained to find the right position of each item. Statistically, analyses were conducted through a one-way ANOVA and Tukey test. The findings of the study revealed that the participants in interactionally modified input group achieved the highest comprehension scores than their peers in other three groups (P<0.05). In addition, modified output group outperformed both premodified input and control groups. However, there was no significant difference between premodified input and control groups and both of them obtained the lowest comprehension scores.