The present study aimed to probe the impact of visual scaffolding using input and output-oriented tasks with different levels of involvement load on Iranian EFL learners' comprehension and production of lexical collocations. For this purpose, 180 male and female intermediate EFL learners were selected and assigned to six experimental groups. Three input-oriented tasks of True-false (load = 1), Matching (load = 2), Multiple-choice (load = 3), and three output-oriented tasks of Short-response (load = 1), Fill-in-the-blanks (load = 2), Sentence formation (load = 3) were developed. All the experimental groups were scaffolded through visual cues. At the end of treatment period, two posttests- a 40-item multiple-choice test and 40-item Fill-in-the-blanks test- were administered to assess the participants' comprehension and production of lexical collocations. To analyze the data, two separate one-way MANOVA procedures were used. The results revealed that visual cues were effective on learners’ collocational achievement. The results also indicated that the output-oriented tasks had a significant positive effect on the comprehension and production of lexical collocations. In addition, tasks with higher involvement load indices were more effective on the comprehension and production of lexical collocations. These finding can have significant pedagogical as well as theoretical implications.