Human activity recognition (HAR) aims to acknowledge activities from a series of observations supported the actions of subjects and therefore the environmental conditions. The HAR research is the basis for many applications including video surveillance, health care, and human-computer interaction (HCI). In this research work, proposed sequential classification problems with evolutionary model and with fuzzy finite state machine. The approach is employed to process and analyze data sets, like activities of daily living (ADL) and activities of daily working (ADW) and later applied with machine learning algorithms namely support vector Classification (SVC), Decision Tree to classify the activities. The technique fuzzy finite state machine and genetic algorithm (G-FFSM) shows better results in terms of performance measures accuracy, precision, recall, f1-scores with 96% average.