AbstractPeak shaving is a common challenge facing most provincial power grids (PPGs) in China. This paper presents an optimization model for the generation scheduling of multiple PPGs using peak-shaving indexes. Loads and power structures were considered to develop four indexes for quantifying the comprehensive peak-shaving needs of a PPG. The peak and off-peak capacity indexes, which represent the upward and downward balance pressures during hours of heavy and light loads, respectively, are functions of the dispatchable power, load extrema, and reserves. The rapid response index is defined to reflect the load-tracking pressure from off-peak to peak hours or vice versa. The fluctuation smoothing index represents the need for flexible regulation in response to frequent load changes in opposite directions. These indexes are normalized and integrated to establish a multiobjective model for peak operations in PPGs, subject to transmission energy and network limitations. This model is transformed into a solvable mixed-integer nonlinear programming model using polynomial and linear approximations and is solved with a branch-and-bound-based method. The model was implemented for the generation scheduling of the Xiluodu Hydropower Plant, which serves two PPGs. Case studies showed that the developed indexes provide accurate guidance for producing day-ahead generation schedules and distributing power among PPGs. The peak-shaving pressures are significantly alleviated in terms of index changes compared to an existing model.