Smoking potent cannabis was linked to 24% of new psychosis cases analysed in a study by King's College London.
The research suggests the risk of psychosis is three times higher for users of potent "skunk-like" cannabis users than non-users.
The study of 780 people was carried out by KCL's Institute of Psychiatry, Psychology and Neuroscience.
A Home Office spokesman said the report underlines the reasons why cannabis is illegal.
Scientists found the risk of psychosis was five times higher for those who use it every day compared with non-users.
They also concluded the use of hash, a milder form of the drug, was not associated with increased risk of psychosis.
Psychosis refers to delusions or hallucinations that can be present in certain psychiatric conditions such as schizophrenia and bipolar disorder.
Risk increased 'threefold'
"Compared with those who had never tried cannabis, users of high potency 'skunk-like' cannabis had a threefold increase in risk of psychosis,' said Dr Marta Di Forti, lead author on the research.
She added: "The results show that psychosis risk in cannabis users depends on both the frequency of use and cannabis potency."
A Home Office spokesman said the findings backed up the government's approach: "Drugs such as cannabis are illegal because scientific and medical evidence demonstrates they are harmful.
"This report serves to emphasise how they can destroy lives and communities."
The spokesman added "there are positive signs our drugs strategy is working", claiming "people going into treatment today are more likely to recover now than in 2010".
Skunk contains more THC than other types of cannabis, which is the main psychoactive ingredient.
Unlike skunk, hashish - cannabis resin - contains substantial quantities of another chemical called cannabidiol or CBD and research suggests this can act as an antidote to the THC, counteracting psychotic side effects.
Sir Robin Murray, professor of psychiatric research at King's, commented: "This paper suggests that we could prevent almost one quarter of cases of psychosis if no-one smoked high potency cannabis.
"This could save young patients a lot of suffering and the NHS a lot of money."
The research was carried out over several years, comparing 410 patients aged 18-65 who reported a first episode of psychosis at a south London psychiatric hospital with 370 healthy participants within the same age range from the same area of London.
It will be published later this week in the Lancet Psychiatry.
The researchers argued that frequency of use and cannabis potency are "essential factors in the mental health effects on users" which are "not sufficiently considered by doctors" at present.
Dr Di Forti called for "a clear public message" to users, comparable to medical advice on alcohol and tobacco.
She said that, as with alcohol, GPs should be encouraged to ask how often and what type of cannabis patients use.
Rosanna O'Connor, director of alcohol, drugs and tobacco at Public Health England, responded: "No drug use is without risk as this report demonstrates.
"Anyone having problems with drug use should seek help from their local specialist drug services. It is important to remember that treatment for all types of drug problems, including cannabis, are readily available and very effective".