<p>Quantum approximate optimization algorithm (QAOA) is used for NP-hard problems on noisy intermediate-scale quantum (NISQ) devices. We demonstrate <u>QAOA's</u> near-optimum maximum likelihood (ML) decoding for short block-lengths in Gaussian channels using random linear codes and channel coded modulation. Simulations with a p-layer QAOA decoder for p ∈ [1, 4], coding rates R = k/n ∈ [0.3, 1], signal-to-noise ratio (SNR) ∈ [0, 10] dB and k ∈ [10, 26] show near-optimum bit and block error rates. We conjecture near-optimum performance for p ∈ [1, 10], R = 0.5, SNR = 10 dB and k ≦ 250 indicating <u>QAOA's</u> potential in short block-length decoding.</p>