Unpleasant code bad smells result from poor design in the first place, where the code is written in a repetitive, incomprehensible, or inefficient manner, which leads to negative consequences that affect the future maintenance activities and subsequent improvements. Examples of code bad smell are large class, duplicate code, long method, long parameter list, etc. There are many reasons for the appearance of code bad smell such as failure to adhere to the principles of object-oriented design, lack of standards, lack of clear guidelines for code design, or lack of experience in programming. To produce high-quality programs, design quality requirements must be adhered to, which positively affects the quality of the product. This research work proposes a hybrid approach using the bat algorithm (BA) and Machine Learning (ML) techniques to automatically generate a detection model to discover the large class code bad smell in object-oriented design. The generated model from this hybrid approach encompasses a pair of a coupling metric and a cohesion metric. The efficiency of the resulting detection model to discover the large class code bad smell is examined by using several ML models offered by WEKA software. The proposed BA-based solution achieved distinctive results, which confirmed the possibility of using BA technology in this field. Keywords: Large Class Bad Smell, Bat Algorithm, Machine Learning