Radio frequency fingerprint identification (RFFI) has gained popularity for physical layer security. Although deep learning-based RFFI has achieved state-of-the-art results, there is still performance degradation in varying wireless environments. To address this challenge, we integrate channel equalization with domain adaptation (DA) to boost the accuracy and practicability of RFFI for devices operating under the IEEE 802.11 standard. Firstly, we apply equalization on raw signals to refine the quality of input for convolutional neural network (CNN) to suppress the channel effects. Then, correlation alignment (CORAL) is used to adapt the CNN for cross-scenario applications, which provides compensation for residual channel effects left by channel equalization. This innovative fusion of techniques not only addresses the inherent limitations of each method when used in isolation but also results in a robust RFFI model capable of adapting to new environments, as substantiated by our simulations and real-world data experiments.