Although the RIRO method based on Retinex decomposition has interpretability, due to the lack of explicit color space constraints in its optimization model, some problems such as color shift and uneven brightness often occur. To overcome these problems, this paper proposes a Retinex-inspired reconstruction optimization model based on bidirectional color-consistency constraints. Specifically, the target image is transformed from the RGB color space to the YCbCr color space, and the consistency constraint is imposed on the transformed YCbCr image. Conversely, after transforming from the YCbCr color space back to the RGB color space, the color consistency constraint is enforced in the RGB color space. Then, this bidirectional color consistency constraint is integrated into the Retinex-based optimization model. Based on the theory of the alternating direction method of multipliers, the model is expanded into an interpretable network framework, which performs joint optimization on the reflectance map, illumination map, YCbCr color components, reconstructed image, and augmented Lagrange multipliers. Experimental results show that, compared with existing state-of-the-art methods, the proposed method has significant advantages in brightness balance, color fidelity and texture detail restoration.