Free-space optical communication (FSOC) systems face susceptibility to several factors, such as transmission distance, atmospheric turbulence, and alignment errors. These elements contribute to fluctuations in the signal strength reaching the receiver. The resultant signal fluctuations can result in misjudgments and an elevated bit error rate (BER). This paper proposes an adaptive decision threshold algorithm based on a sliding window (ADTSW). By estimating received signal parameters and delimiting the amplitude interval, the algorithm ensures that the decision threshold tracks signal fluctuations, thereby reducing signal misjudgment. The effectiveness of the algorithm is validated through simulations and experimentation. When the signal peak-to-peak value fluctuates, simulation results demonstrate that the proposed algorithm achieves a 1-order-of-magnitude reduction in BER compared to the traditional fixed decision threshold (FDT) method. Under the influence of weak atmospheric turbulence with different scintillation variance, both simulation and experimentation indicate a 1-order-of-magnitude reduction in BER compared to the FDT method. The ADTSW algorithm proves its capability in minimizing misjudgments, thereby effectively reducing BER and improving communication quality. © 2024 Optica Publishing Group © 2024 Optica Publishing Group (formerly OSA). All rights reserved.