To construct a multi‐energy complementary system for intelligent building group, a direct control strategy for central air conditioning based on discrete Fourier transform is proposed. This strategy can reduce the deviation between actual load and transaction results while absorbing photovoltaic (PV) power generation and improving the enterprise economy in the power market environment. The strategy uses the discrete Fourier transform (DFT) to find the scheduling demand of the central air‐conditioning (AC) group and realizes the scheduling demand by changing the temperature setting value. Also, the strategy can assess AC status in real‐time, which combines with ultra‐short‐term load forecasting to adjust future purchase planned values. The simulation results show that the strategy can realize a real‐time load track with the power purchase plan while absorbing PV power generation with low action times and temperature changes, which improves the absorption of distributed PV power generation and reduces the assessment bias.