A prediction for any sport type is something we always try in our sport life. An early prediction is always helpful for the team management to work on their plan and train the team quickly. The same is true for football game also. Football is one of the most difficult sport to predict the result. Currently, the statistical modelling approach is applying for prediction of the match result. However, the statistical approach predicts the soccer result with poor accuracy. in addition to this, the success of the team is depend on many various factors, which might leads to insufficient prediction and inconsistence results. Furthermore, the evaluation of a player performance for the squad formation, transfer and strategic planning is very important because of the player and team performance is essential for both market value of a player and prediction of the soccer result. Both the factors have majorly depended on a various attribute of the players. In this thesis we proposed a machine learning approach prediction model development for the success and consistency of football games. A Machine learning algorithms assists the coaches and managers in result prediction, player performance assessment, player injury prediction, sport talent identification and game strategy evaluation. This thesis is primarily focusing on developing a predictive model that provide an appropriate relationship between the attribute value, performance value and the result of the team in the premier league estimated using supervised machine learning algorithms by reviewing various related paper and existing system as a methodology and by using support vector machine and K nearest neighbor algorithm to develop a predictive model and also using recursive feature elimination or extra tree classifier for feature selection and ranking of features respectively. The popular algorithm support vector machine provides a better accuracy than the other algorithms. We have got the accuracy of 77% and 68% when we implement all the fourteen features using supervised machine learning and K nearest neighbor respectively. After removing the feature having less influence on the outcome of the match using recursive feature elimination algorithm the accuracy become 70% and 60% for supervised machine learning and K nearest neighbor algorithms respectively. When we implement all the features without removing the feature having less influence on the outcome of the match provide better accuracy. This helps the team coach, team managements, supporters and online funs to identify the future prospects in the game of football. Keyword: Football, K Nearest Neighbor, Recursive Feature Elimination, Soccer Result Prediction, Support Vector Machine