Science, technology, engineering and mathematics (STEM) competencies are becoming increasingly important in times of rapid technological innovation and global challenges. This approach increases students' personal scientific literacy, including international economic competitiveness and ethical citizenship. STEM education has been actively developing in Kazakhstan since 2014. Experimental schools are even being created in the country to develop and test new teaching methods. However, recent results from the International Student Assessment Program show that even in economically developed countries such as Europe, the US and Australia, around 20 percent of students lack sufficient skills in math or science. This trend highlights the urgent need for action on STEM education. Although mathematics is widely recognized as the foundation of all other STEM subjects, there is strong evidence that its role in integrated STEM education is underestimated. Mathematics is important for many careers, especially science, technology and engineering. However, mathematics is often perceived as difficult and, as a result, many students drop out of science, technology, engineering and mathematics (STEM) subjects and do not choose careers in science, technology and engineering. If the mental load in the classroom increases, students' interest in the subject also diminishes. The tasks related to real-life situations arouse the children's interest and desire to solve problems, as these skills can be useful in real life in the future. The tasks related to real-life situations arouse the children's interest and desire to solve tasks, as these skills can be useful in real life in the future. In this paper we propose an alternative approach to changes in mathematics education and show how this alternative applies to STEM education. We also address an element of this problem by exploring the role of mathematics in STEM education and how it can be improved through three interdisciplinary approaches: (1) twenty-first century skills; (2) mathematical modelling. At the end of the paper, we discuss the potential for research in relation to two aspects and indicate what work needs to be done in the future.