The current scenario requires the application of new computational tools for the feed formulation strategy that uses mathematical modeling in decision making. Noteworthy is the nonlinear programming, which aims not only to formulate a diet that meets the needs of the animal, but also the minimum cost and the maximum profit margin. Thus, the work aimed to validate the use of the nonlinear model (NLM), with maximization of the economic return, through estimates of animal performance and feed costs, according to the price variation of the kg of the broiler (price historical average of 2009 and 2010), the phases of creation and sex. For this purpose, 480 broiler broiler chickens, 240 males and 240 females of the same strain (Cobb 500) were used, from 1 to 56 days of age. The experimental design was entirely randomized, totaling 6 treatments (increasing or decreasing the average historical price of live chicken by 25% or 50%), with 4 replicates and 10 broiler chickens per experimental plot. Performance (weight gain and feed consumption), total energy consumption and profit margin were evaluated. Regarding the formulation principle (Linear and Nonlinear), the performance was very similar in relation to the studied parameters. However, when simulated values of 50% below the historical average, performance was significantly impaired in this specific condition. However, due to the profit margin, it demonstrated that the principle of nonlinear formulation allows to significantly reduce losses (P <0.05), mainly in unfavorable conditions of the price of chicken in the market. It is concluded that the nonlinear principle is more appropriate, since the requirements of all nutrients are automatically adjusted by the mathematical model and with the premise of increasing profitability, different from the linear one, which is to achieve maximum performance and not is directly related to the economic factor.