This paper suggests a stochastic volatility term structure model applied to the pricing of electricity swaptions in the Nordic power market traded at the Nasdaq OMX Commodities exchange. The volatility structure in the model is specified as a product of a time-dependent function, which handles the maturity effect, and a Cox-Ingersoll-Ross process for the stochastic volatility. We employ a Fourier-based approach to pricing electricity swaptions and perform an empirical analysis by calibrating the model to a data set consisting of more than 12 000 pairs of implied bid-ask volatilities, corresponding to swaption prices from the Nordic power market. We show that our model outperforms the lognormal benchmark both in- and out-of-sample.