To address the issues of the optical transmission model being incomplete and the reconstruction algorithm parameters being unclear in passive non-line-of-sight (NLOS) target reconstruction, this study proposes an optical transmission model based on BRDF and a NIQE-integrated FISTA algorithm. The optical transmission model for passive NLOS scenarios is established based on BRDF, discretized, and formulated into an objective function comprising data fidelity and smoothness terms. NIQE is employed as an indicator to integrate with the FISTA iterative algorithm to solve the optimization problem, yielding the optimal reconstruction results. In the reconstruction of the MNIST handwritten digit dataset, the PSNR reached 11.34 dB, and the SSIM was 0.79. Compared to polarization-enhanced methods, the PSNR of the reconstructed target improved by 96.92%, and the SSIM increased by 83.72%. Compared to deep learning methods, the PSNR increased by 74.26%, and the SSIM reached 95.18%. This approach effectively enhances the performance of passive NLOS target reconstruction.