Wireless sensor networks (WSNs) have turned up as a promising technology due to their deployment nature and the knock-off feature of cost-effectiveness. The rapid increase in the demand for wireless applications has opened up new security vulnerabilities which are related to their infrastructure-less nature and their nature of transmission and hence made WSNs the centre point of attraction for attackers and intruders. Machine Learning has embedded comprehensive solutions to accord with modern security breaches. It has gained a mark of deterrent technology by shoring up the infrastructure of security through mapping security breaches and performing the identification of unknown patterns. This chapter focuses on the illustration of security breaches in wireless sensor networks along with the glimpse of reduction in attacks through dynamic algorithms and strategies of Machine Learning. We have proposed some feature selection methodologies in order to identify the best features out of the available network dataset. These feature selection methods are evaluated against the existing machine learning classifiers in order to identify the best feature selection strategy and the best classifier.