A report by three MIT researchers was presented at the international joint conference on artificial intelligence in Argentina in July.
In an article on the MIT News website, the lead author of the report, Kalyan Veeramachaneni, said the wind industry had been using a simplistic technique to measure predicted wind speeds.
Using data from a nearby weather station and an anemometer at the proposed project site, consultants use a bell curve model to predict wind speeds.
This typically takes 12 months to achieve reliable enough data to site a wind farm.
Veermachaneni said this technique was "not an accurate representation of the data".
The new model, developed by Veeramcheneni and his team, uses data from several weather stations in the area and a different algorithm to predict wind speeds with greater accuracy and less historical data.
The researchers applied their model to historical data gathered by a wind industry consultant at a project site.
Veeramcheneni's team found using just three months of data, they could "predict wind speeds over the next two years three times more accurately than existing models could do with eight months of data".
With revisions to the technique, it is believed the model could be up twice as accurate than originally predicted.