In the Swedish study, mice whose water contained 10% alcohol had a lower risk of rheumatoid arthritis.
The team said it was not possible to say exactly how much alcohol would have the same effect for humans.
But UK arthritis experts cast doubt on the relevance of the Proceedings of the National Academy of Sciences study findings for treating human disease.
Low or moderate alcohol intake has been shown to benefit people in a number of ways, such as lowering the risk of heart disease.
But drinking too much causes complications including liver damage.
Translation 'not possible'
In the study carried out by a team at Gothenburg University, mice were given injections of collagen to induce rheumatoid arthritis (RA).
They were then either given untreated tap water, or water with 10% pure alcohol - a level which was not toxic to the livers - for between six and eight weeks.
RA was found to develop significantly more slowly in the mice given the alcohol, and had less severe symptoms once the disease did start to progress.
The researchers say alcohol may boost production of the male hormone testosterone.
This then restricts a key part of the mechanism which releases proteins called cytokines which cause inflammation.
The team also found that acetaldehyde, which is formed when the body processes alcohol, can produce similar protective effects.
Professor Andrzej Tarkowski, who led the research, said: "We can't translate these results to find out the therapeutic dose in humans.
"The mice were given a dose of 10% of alcohol in their water, but we don't know if it would be the same for humans. It would probably be lower."
He added: "One possibility would be to use acetaldehyde, which produced similar effects, but which could provide an alternative non-addictive treatment."
Testosterone 'offers protection'
Professor Alan Silman, incoming medical director at the Arthritis Research Campaign said: "Studies in humans have not shown any relationship either protective or in terms of increased risk between alcohol intake and rheumatoid arthritis.
"There is no really close mouse model of rheumatoid arthritis, and the collagen-induced arthritis may not be a particularly good human model."
But he added: "There is evidence to suggest that alcohol can increase testosterone and increased testosterone may protect against development of rheumatoid arthritis.
"And rheumatoid arthritis is rare in younger males and, at all ages, is more common in females.
"It is possible therefore that in this mouse model, alcohol may have had some effect in relation to arthritis.
"However it is doubtful whether this would have much influence in the human situation. "
