Estimating the software work is a crucial job of persons participating in software project management. The difficulty in predicting effort is compounded by the fact that software development is always changing. Several techniques for estimating software development costs have been developed over the last three decades. There are a variety of cost estimation methodologies, algorithmic models, non-algorithmic models, and machine learning methods to choose from. To improve accuracy, machine learning approaches are combined with algorithmic or non-algorithmic models. Researchers in past worked on the effort and time estimation by using one type of development methodology in their work. Currently, the companies uses both agile and traditional techniques to software development. A comparison of agile and traditional development utilizing the neural network (NN) and genetic algorithm is presented in this research (GA). Estimation is performed on the Zia dataset and a github dataset using story points and lines of code, respectively. The smallest error and highest accuracy were attained utilizing machine learning approaches for projected effort values. The value of R2 based on story point is achieved using neural network and genetic algorithm is 0.97 and 0.96 respectively. On other hand, the value of R2 based on lines of code is achieved using neural network and genetic algorithm is 0.94 and 0.80 respectively. The mean magnitude relative error is used for comparison of proposed models with previous works. The dataset with the story point give best results followed by projects with lines of code.