People of any age or gender might be affected by diabetes, an autoimmune disease. It is a prevalent ailment that impacts the essential body parts. A survey conducted over the past 20 years shows that the number of diabetics has tripled. According to the IDF, currently, around 382 million people have diabetes; it may be expected to exceed up to 592 million people by 2035. The main interest of this project is to utilize and choose the most appropriate model by using physiological pnarameters to predict a person’s risk of diabetes. We analyze the diabetes detection algorithms created by several scientists in this work. Most research is based on the PIMA Indian Diabetes Dataset (Pregnancy-related Diabetes). The scope of machine learning algorithms, like Decision Trees (DT), Naive Bayes (NB), Support Vector Machines (SVM), Logistic Regression (LR), and Random Forest algorithms, are used to detect diabetes in the period of pregnancy. The purpose of using a variety of algorithms is to assess how well commonly used machine learning algorithms work and determine which algorithm would be most appropriate for our needs. For SVM, an accuracy of 77.9%; for DT, an accuracy of 94.14%; for LR, an accuracy of 78.2%; for NB, an accuracy of 76.3%; and for RF algorithm, 72.7% accuracy was obtained.