The high proportion of distributed PV generation connected to a distribution network results in a nodal voltage over limit problem. Thus a distributed PV reactive power regulation method based on affine adjustable robust optimization is proposed in this paper. First, a comprehensive optimization model is established to minimize the active power loss and nodal voltage deviation of the distribution network, and the polyhedral linearization method is used to transform the comprehensive optimization model into a quadratic programming model. Then, considering the budget of uncertainty of active power of distributed PV, the robust optimization model is transformed into a deterministic quadratic programming model through dual transformation. The simulation results on IEEE-33 and IEEE-123 buses distribution systems with high proportion of distributed PV show that, compared with zero reactive power regulation of distributed PV, the proposed affine adjustable robust optimization method can greatly improve the optimization results of network loss and nodal voltage over that of the traditional robust optimization method and has lower conservatism, which is closest to the results of Monte Carlo simulation. This work is supported by the Science and Technology Plan Project of Hebei Province (No. 19212102D) and the Science and Technology Project of State Grid Hebei Electric Power Co., Ltd. (No. kj2020-049).