<p>Software defects have always been crucial for software developers. Defects cause a software to show illicit behaviour and harness the quality of the software and ultimately posing a critical challenge for the developers. Code smells may indicate the presence of defects which may affect the maintenance of the software adversely. Several tools and techniques have been proposed for identifying and classifying code smells in the source code. Most of them are based on single label classifier and traditional approaches. This study aims to explore various multilabel classification (MLC) algorithms and base classifiers which can be used for the detection of code smells. Six different multilabel classification methods in combination with twelve base classifiers have been applied on two sets of multilabel code smell datasets at class level and method level and a comparative study among them has been proposed. This study also found that there are several other significant multilabel classfication algorithms such as Conditional Dependency Network (CDN), Four class Pairwise (FW) other than widely used algorithms such as Binary Relevance (BR) and Classifier Chain (CC) or Label Combination (LC) that has not been addressed earlier, can also give promising results. It is also found that apart from widely used Random Forest base classifier, there exist other base classifiers such as Attribute Selected Classifier (ASC), Iterative Classifier Optimizer (ICO), LogitBoost etc. which gives the same or even better results than Random Forest. The study suggests that multilabel classification is a potential technique in the realm of classification which can be considered and explored more by the researchers working in the area of software defects.</p>