Two experiments were conducted to investigate whether articulatory suppression and memory updating affect phonological similarity effect. In the first experiment, 24 subjects were tested for immediate serial recall of visually presented letter sequences in which letters were either phonologically similar or dissimilar within each sequence. Each block of the letter sequences was to be remembered under silent reading or articulatory suppression conditions. The results showed the advantage of the dissimilar condition over the similar condition in the silent reading condition. On the other hand, this phonological similarity effect was not shown in the articulatory suppression condition. Twelve subjects participated in the second experiment on the effect of memory updating. Memory updating reduced the recall performance, but it did not affect the phonological similarity effect. These results were interpreted in terms of the working memory model.