Nowadays, the persuit of prediction of chemical bond formation among researchers is evident to discover new drugs or for new discovery of chemicals. Predictive modeling gives statistics to predict outcomes. Most of time the event one wants to predict is in the future, but predictive modeling applicable to any type of event, regardless of when it occurred. In this paper, we focused on comparative analysis of various Prediction algorithms to estimate the best algorithm for the prediction of chemical bond formation with observations. The performances of these techniques are compared and it is observed that parallel Genetic Algorithm provides better performance results in accuracy and speedup as compared to other prediction techniques on GPU.