With its unique advantages in processing massive and complex data, machine learning has opened up new avenues to break through the existing bottlenecks in the treatment and prognostic prediction of IgA nephropathy. In terms of treatment decision-making, reinforcement learning and multi-task learning are capable of simulating the long-term efficacy of different clinical intervention plans, and accurately balance the therapeutic benefits and potential risks. In terms of efficacy evaluation, supervised learning and graph neural networks can effectively screen hormone-sensitive biomarkers and explore the key molecular pathways of IgA nephropathy. In terms of prognosis prediction, deep learning survival analysis models such as ensemble learning and DeepSurv can achieve refined risk stratification for IgA nephropathy, while temporal models including Transformer can dynamically track changes in renal function. This paper reviews the latest research progress of machine learning in the precise treatment of IgA nephropathy, as well as its application status in fields such as pathological image analysis and mobile health platforms, aiming to promote the transformation of IgA nephropathy treatment from traditional empirical therapy to a data-driven precision treatment mode.