Abstract: In a multi-market environment, electricity retailers need to allocate the proportion of electricity purchased from the contract market, the day-ahead market and the real-time market in order to reduce risks and increase profits.This paper considers the implementation of peak-to-valley electricity prices on the side of power sales, studies the revenue space of electricity procurement and sale for electricity retailers, and uses CVaR to quantify the losses that electricity retailers may suffer;Since CVaR mainly reflects the size of loss, it can not control the probability of the portfolio purchasing loss of electricity retailers. Therefore, the skewness constraint is introduced into the portfolio purchasing optimization model, and the probability of power purchase loss of the electricity retailers are controlled by controlling the degree of skewness.The greater the skewness, the lower the probability of a low rate of return and the greater the probability of a high rate of return.Taking the maximization of sales revenue of electricity retailers as a target function, the skewness-CVaR based portfolio purchasing optimization model for electricity retailers under peak - valley prices is established.The example results show that this model provides new risk measurement indicators and power purchase plans for electricity retailers.