Spatial frequency adaptive processing (SFAP) is the dimension-reduced alternative of spatial temporal adaptive processing (STAP), which can achieve good interference mitigation capability with less computational cost. However, just as STAP, the biases of the ranging code phase and carrier phase induced by the adaptive array restrict SFAP from being applied in high-precision global navigation satellite system receivers. To address this issue, a new data-oriented calibration method is proposed. Unlike previous bias reduction methods, which compensate the measurement biases with estimated values in the tracking loop, this method carries out calibration directly on the received data in the frequency domain, thus avoids modification to the rest parts of the receiver other than the SFAP module. The calibration value in each frequency bin is determined from the transfer function of the SFAP array. Simulations are performed on a practical array antenna designed by the high-frequency structure simulator software with high fidelity, and the results verify the effectiveness of this method.