This master's thesis explores the application of inertial sensors in smartphones within the context of step tracking, with a particular emphasis on the accelerometer as a pivotal sensor for precise measurements. Other pertinent sensors such as motion, environmental, and positional sensors are analyzed, investigating their mutual interaction. The latest algorithms for step detection are examined, encompassing methods based on parameters, machine learning, and deep learning. The objective is to analyze the advantages and limitations of each method, with a focus on selecting the most optimal approach for implementation in the application. The process of implementing the Android application is comprehensively described, guided by pertinent scientific articles. Key steps are illustrated, including the peak detection method and five data processing stages, ranging from linear interpolation to selecting accurate peaks through periodicity and similarity analysis. Testing involving ten participants covered various activities and device positions, with results compared to popular applications like "Mi Fitness" and "Pedometer App". Statistical analysis provides insights into accuracy, the influence of testing conditions, and the overall reliability of the developed step-counting system.