Human voice is basically sound which is made by humans from their vocal tracts. Voice is made of different constituents and has various characteristics such as frequency, amplitude etc. These characteristics are produced by combination of vocal folds and articulations. This paper reflects development of a system using these characteristics which altogether are called acoustic parameters to detect the gender of the speaker. We have used four models to classify the genders namely CART, XGBoost, SVM and Random Forest. An ensemble of all the models is also used to make the entire system more accurate. This system can be used as a building block for many other softwares where it will take the first step to extract the acoustic parameters and detect the gender of the speaker.