Abstract: In this study we have analyzed basics of software measurements which is indicated by software defect prediction. We have primarily focused on static code measures to find the probability of defect. These measures have been used to estimate the accuracy of the algorithms so that we could predict defect, probability of detection and false alarm based on the PROMISE Software Engineering Repository data set. To do this we have applied k-means clustering and Euclidean distance techniques for probability. This analysis helps us to estimate which attribute when chosen alone give more accuracy then others so that they can be used further for more rigorous assessment by some other means.