Offers on the day-ahead market are delivered significantly before the actual time of delivery/reception of electricity, which significantly complicates planning and increases costs for stochastic sources of electricity such as wind farms. The notification of the wind farm production plan up to 36 hours before the actual moment of electricity delivery to the system (in systems without intra-day markets) results in significant deviations from the defined exchange plan and consequently high balancing costs. In this paper, a description of the method for determining the optimal wind farm offers on the market one day in advance is given in order to achieve the maximum expected profit for wind farms. Defining the wind farm production plan in an amount equal to the expected production of the wind farm sometimes does not guarantee maximum profit. For example, in the case when the difference between the price of balancing energy to cover deviations with a positive sign and the market price is significantly greater than the difference between the price of balancing energy to cover deviations with a negative sign and the market price, it is optimal for the wind farm to define a production plan in the amount below expected production. With this strategy, the wind farm is less exposed to high balancing cost in case of positive deviation. In the opposite case, when the difference between the price of the balancing energy to cover the deviation with a negative sign and the market price is significantly higher, it is optimal for the wind farm to define a production plan in the amount above the expected production. This strategy avoids the sale of surplus energy compared to the planned one at prices significantly lower than the market price. With this in mind, it is obvious that the application of the method makes sense in cases where the balancing energy prices are unevenly distributed, i.e. in cases where the wind farm production forecast is unevenly distributed.