As Alves (2006), ‘ ‘ esteem software cost is not a task easy, therefore it involves many 0 variable as the generated product quality, the established stated periods, the involved human resources, the identified risks, of among others same relevncia’ ‘. However, the organizations need necessarily esteem the involved effort and the cost in the process of software development. In this direction, Pressman (2006) presents some options to obtain estimates of trustworthy cost and effort: 1.Adie the estimate until the project is more in advance; 2.Baseie the estimates in similar projects, that already had been completed; 3.Use techniques of relatively simple decomposition to generate estimates of costs and effort of the project; 4.Use one or more empirical models for estimate of cost and effort of software. With same objective, to obtain estimates of trustworthy cost and effort, Sommerville (2007) presents some below techniques of estimates of costs of software, as described: 3.Estimativa algorithmic 1.Modelagem of costs 2.Julgamento of specialists for analogy 4.Lei of Parkinson 5.Atribuio of price to gain One or more described techniques for Pressman (2006) and Sommerville (2007) can (m) be used (s) to produce estimates of costs. Each technique presents positive points and negative, to get itself better resulted it is necessary to use one technique more than and to compare its results. In last the two decades, many studies have been carried through in the area of estimates of costs for software projects. Amongst the techniques of cited estimates the present article approaches two estimates of cost and effort according to Pressman (2006): Empirical techniques of Decomposition and Models. The boarding Empirical Models also is seen according to Sommerville (2007), nominated as algorithmic Modeling of costs.