Accurate state of health (SOH) estimation is important for the safe operation of lithium-ion batteries. Existing supervised learning methods rely on large amounts of labeled data and may not fully exploit degradation information in battery curves. To address this problem, this paper proposes a masked pre-training method based on incremental capacity (IC) curves for SOH estimation. The IC curves are divided into sequential patches, and a proportion of the patches are randomly masked during self-supervised pre-training. By reconstructing the masked patches from the remaining curve information, the model learns useful representations from unlabeled IC curves. The learned representations are further used for SOH estimation through fine-tuning. The influence of the masking ratio is investigated, and comparative experiments using IC, voltage, and current curves as masking inputs, together with different self-supervised reconstruction methods, verify the effectiveness of the proposed method.Accurate state of health (SOH) estimation is important for the safe operation of lithium-ion batteries. Existing supervised learning methods rely on large amounts of labeled data and may not fully exploit degradation information in battery curves. To address this problem, this paper proposes a masked pre-training method based on incremental capacity (IC) curves for SOH estimation. The IC curves are divided into sequential patches, and a proportion of the patches are randomly masked during self-supervised pre-training. By reconstructing the masked patches from the remaining curve information, the model learns useful representations from unlabeled IC curves. The learned representations are further used for SOH estimation through fine-tuning. The influence of the masking ratio is investigated, and comparative experiments using IC, voltage, and current curves as masking inputs, together with different self-supervised reconstruction methods, verify the effectiveness of the proposed method.