The coupled space-time correlation among output series in multiple wind farms has great effects on the operation of power systems. Space-time auto regressive moving average (ST-ARMA) model can restore statistical model of coupled space-time correlation of multi-dimensional sequences in a relatively simple form. In order to build an effective model for multi-dimensional wind power series

the measured wind power data from the perspective of space-time series are statistically analyzed

especially for the coupled space-time correlation. Then

a spatial relation matrix is used to describe the location of each wind farm

and embedded into the auto regression process of ST-ARMA model in order to simulate the coupled space-time correlation of wind power series. The ST-ARMA model has restored not only the temporal and spatial correlation of original wind power series

but also the coupled characteristic between the two kinds of correlation. This model can be used to produce huge amount of simulation data with the same characteristics of actual wind power output

which is able to provide data basics for the planning and operation research of wind power integration.