In this paper, we propose an efficient symbol detection algorithm for space-frequency OFDM (SF-OFDM) transmit diversity scheme and present the implementation results of the SF-OFDM wireless LAN (WLAN) baseband processor with the proposed algorithm. When the number of sub-carriers in SF-OFDM scheme is small, the interference between adjacent sub-carriers is generated. The proposed algorithm eliminates this interference in a parallel manner and obtains a considerable performance improvement over the conventional detection algorithm. The bit error rate (BER) performance of the proposed detection algorithm is evaluated by the simulation. In the case of 2 transmit and 2 receive antennas, at BER=10-4 the proposed algorithm obtains about 3 dB gain over the conventional detection algorithm. The packet error rate (PER), link throughput, and coverage performance of the SF-OFDM WLAN with the proposed detection algorithm are also estimated. For the target throughput at 80% of the peak data rate, the SF-OFDM WLAN achieves the average SNR gain of about 5.95 dB and the average coverage gain of 3.98 meter. The SF-OFDM WLAN baseband processor with the proposed algorithm was designed in a hardware description language and synthesized to gate-level circuits using 0.18 μm 1.8 V CMOS standard cell library. With the division-free architecture, the total logic gate count for the processor is 945 K. The real-time operation is verified and evaluated using a FPGA test system.