Abstract: More distinguishable and consistent pixel features for each category will benefit the semantic segmentation under various settings.Existing efforts to mine better pixel-level features attempt to explicitly model the categorical distribution, which fails to achieve optimal due to the significant pixel feature variance.Moreover, prior research endeavors have scarcely delved into the thorough analysis and meticulous handling of pixel-level variances, leaving semantic segmentation at a coarse granularity.In this work, We analyze the causes of pixel-level variance and raise the concept of $\textbf{pixel learning}$ to concentrate on the tailored learning process of pixels, handle pixel-level variance, and enhance the segmentation model's per-pixel recognition capability.Under the context of the pixel learning scheme, each image is viewed as a distribution of pixels, and pixel learning aims to pursue consistent pixel representation inside an image, continuously align pixels from different images (distributions), and eventually achieve consistent pixel representation for each category.We proposed a pure pixel-level learning framework, namely PiXL, which consists of a pixel partition module to divide pixels into sub-domains, a prototype generation, a selection module to prepare targets for subsequent alignment, and a pixel alignment module to guarantee pixel feature consistency intra- and inter-images.Extensive evaluations of multiple learning paradigms, including unsupervised domain adaptation and semi-/fully-supervised segmentation, show that PiXL outperforms state-of-the-art performances, especially when annotated images are scarce.Visualization of the embedding space further demonstrates that pixel learning attains a superior representation of pixel features.The code will be available upon acceptance.