Researchers at three British universities, who looked at data of about 93,000 people, also found that tricyclic antidepressants also reduce the risk of glioma -- the most common type of brain cancer -- by up to two-thirds.
Taking larger doses for longer increases the preventative effect, the researchers said.
Although the results are startling, it is highly unlikely such drugs, many of them are sedatives, would be widely prescribed to those without mental health problems because of their side-effects.
However, the researchers are excited because they say people at a genetically higher risk of the two types of cancer could be prescribed them, the Daily Telegraph reported.
The finding, published in the British Journal of Cancer, could also lead to the development of specifically designed pharmaceuticals to tackle bowel and brain cancer, said Dr Tim Bates of Lincoln University.
He explained that tricyclic antidepressants worked by attacking the "Achilles' heel" of some cancer cells, or their mitochondria that enable them to function.
"As cancer mitochondria are biochemically different from mitochondria in normal non-cancer cells, they represent an Achilles' heel," he said.
Tricyclic antidepressants appeared to interfere with the normal working of mitochondria in bowel and glioma cancer cells, he said.
"The cancer prevention action of these drugs may translate into one that is also useful in treating glioma, both in adults and in children, and colorectal cancer," he added.
The study compared about 31,500 people with cancer with about 61,500 people without, and was adjusted for age, gender, smoking, obesity and other factors.
It showed that people on tricyclic antidepressants were between 16 and 21 per cent less likely to have developed bowel cancer, with those who had been taking them at higher doses for longer receiving greater protection.
Bowel cancer is the second bigger cancer killer after lung cancer, which kills 16,000 people a year in the UK alone.
For glioma, tumours of the brain and spine, which kill up to 2,000 a year in Britain, the reduced risk was between 41 and 64 per cent. There was no effect on reducing incidence of other types of cancer.
