Extensive research has been conducted to enhance the accuracy of indoor person identification and support context-aware home services. However, existing methods often struggle with low classification performance, particularly in terms of correct classification rate (CCR), due to various technical challenges. This study introduces an advanced system for identifying individuals in smart homes by integrating pyroelectric infrared (PIR) sensors with floor-pressure sensors. A cooperative multi-sensor approach is employed to derive explicit information about a person’s body size, thereby improving identification accuracy. Additionally, a novel machine learning-based strategy is implemented to extract a unique feature vector representing body size. To further enhance CCR, neural networks (NN) and support vector machines (SVM) are utilized. A prototype of the system was developed and tested through multiple scenarios, demonstrating its effectiveness in achieving high CCR values.