The present invention discloses an effective workday construction method for a monthly electricity consumption prediction model. The method comprises the following steps: dividing the electricity consumption during the Spring Festival into three periods of time, setting an adjustment coefficient for the first seven days of the Spring Festival as c1, that for seven days during the Spring Festival as c2 and that of the last seven days of the Spring Festival as c3 to obtain a formula as shown in the specification, and calculating c1, c2 and C3 three numerical values as initial values of the adjustment coefficients; and determining a scope for step length searching adjustment coefficients, converting the number of effective workdays in the month of the Spring Festival, further calculating an average daily electricity sale set as input data of the electricity consumption prediction model to perform prediction, and according to prediction errors of the model, determining the optimum adjustment coefficients c1', c2' and c3', thereby determining the effective workdays. A predicted value of the average daily electricity sale, calculated by virtue of the model, is multiplied by the effective workdays to obtain a predicted value of monthly data. According to the method disclosed by the present invention, inaccurate data caused by fluctuation of the date of the Spring Festival appearing in a calendar month are corrected, so that a mobile holiday effect is caught to ensure that data accounting is accurate.