The paper in attachment might help you

Meth users who want to break their addiction can join an AA group. These groups can be very supportive, and sometimes it works well. (However, it is not my field of expertise.)

NIAAA has a Cognitive Behavorial Treatment approach which I have use and in combination with CBT techniques from David Burns, M.D., PhD.

Saw a sobering study description - "...a high proportion of meth users...have psychotic disorders, the majority of which are substance–induced."

Can any sort of language-based "cognitive" stuff be done with psychosis!?

Not really.  Methamphetamine is a very habituating drug and support is the only help that i am aware of

There is a dearth of outcome data for all treatment modalities for all substance use disorders.  That said, the general consensus of clinicians IN THE UNITED STATES is against substitution (medication) therapy for stimulants (cocaine, methamphetamine, methylphenidate, d-amphetimine, a-amphetimine, dextroamphetamine, e.g. all DOPAMINE upregulators).  [In response to Dr. Rich's comment, I think the original question was aimed at stopping individuals from using methamphetamine using non-medicinal means.  Yes methamphetamine can induce psychosis, but clinicians can stop this not by treating psychosis directly but by trying to get meth users to stop using it.]  So, for clinicians wanting to take an abstinence-based approach, it seems that "talk therapy" like CBT and support-group therapy like 12-step programs (NA, AA, CA, etc) have shown success - but only anecdotally - we don't really know how effective these modalities are and for which patients they work, and how well they work, in achieving long-term remission from methamphetamine (stimulant) use disorders.  I think that, like with all substance use disorders, the best results will be obtained by keeping patients in long-term outpatient treatment following intensive residential treatment.  As Dr. Buttfield said "support" is the key, whatever the particular structure of a pt's aftercare plan.  Given how little real outcome data we have, the best strategy is probably to apply all available treatment modalities and hope that something helps.  But the *anecdotal* results so far are not terribly encouraging; most pts with stimulant use disorders have great difficulty achieving remission for any meaningful period of time.  In other words, the average Px for a pt presenting with stimulant use disorder is not terribly auspicious.  

For now, it seems that the only two viable non-medicinal treatments are direct therapy (CBT or otherwise) and mutual support therapy (12 step or other group therapy approaches).  I really hope that more treatment modalities will be developed for patients suffering from stimulant use disorder.  Remember also that most SUD patients have high comorbidity with other psychiatric disorders (which contribute to the difficulty of achieving absence/remission) -- it is important to Dx and Tx these as well as SUD itself.  

A final comment on the Karila et al article cited above, which reviews medicinal/pharmacological Tx (Rx) treatments.  There are really three types of strategies in medication, and I think it is important to categorize potential medications as such (1) "blockers" - e.g. medications that stop the stimulant from acting, such as naltrexone.  The problem with naltrexone, which works very well with both opioids and alcohol and has shown promise with stimulants, is PATIENT NON-COMPLIANCE.  Patients may take the medication during a study, but often will stop in conjunction with resuming use.  As such, it is not really a treatment, except in long-acting forms such as the monthly IV injection of naltrexone (Vivitrol in the US, manufactured by Alkermes).  The problem with this, again, is that patients will often decline to take medication that blocks the psychoactive effects of the substance they want to use.  Blocking strategies have had weak results so far EXCEPT in conjunction with rigorous aftercare programs, such as those required of physicians in recovery and pharmacists in recovery.

 (2) "Mitigating medications".  In my opinion, these show the most promise.  Bupropion, for example, is a very WEAK stimulant (and antidepressant and anxiolytic in some patients) - and if it can provide enough dopamine/norepinephrine/adrenaline upregulation that a patient will resist the urge to use a much stronger stimulant, then it is BY FAR a better alternative 

(3) substitution (harm reduction).  As Karila et al note, substitution with d-amphetamine, a STRONG stimulant, logically shows promise.  However, as with any substitution strategy, the patient remains dependent on a substitute stimulant with dangerous morbidity - but which may arguably be substantially less harmful than dependence on methamphetamine.  The Karila review, written in 2010, concludes "Despite the lack of success in most studies to date, increasing efforts are being made to develop medications for the treatment of methamphetamine dependence and several promising agents are targets of further research."  I am unaware of any such compounds emerging as likely candidates for medicinal treatment in the past several years.  

Furthermore, LONG-TERM COMPLIANCE is something that is rarely studied in the RCTs cited by Karila.  Unfortunately, there seems to be an inverse relationship between the effectiveness (measured in terms of pt quality of life) and pt compliance on medication.  With a blocking strategy, you get LOW LONG TERM COMPLIANCE, and with a strong-substitution strategy, you get much higher compliance (naturally) but then you are really just switching the patient from methamphetamine to a stimulant with less deleterious effects, while you try to keep the dose at a steady level (very difficult given rapid tolerance and relatively flat dose-response curves of many stimulant-substitution alternatives).  I am personally (subjectively) much more optimistic about opioid substitution therapy than I am about stimulant substation therapy for that reason -- but I have no data to  back up this conjecture.