In the field of wireless health monitoring for specially old age people cognitive wireless sensors which can monitor the daily activities have been reported to be of great importance. Usually a large number of sensors like motion sensor, current sensor, force sensor and flow sensors are widely employed. However there are redundancies in the system and it is required to reduce the number of sensor nodes to lower the electrical load and cost. In this paper we have tried to optimize the selection of the sensor nodes through a fuzzy decision logic based on performance parameters like ease of installation, power consumption, cost and occurrence of false signals. It has been observed that the performance of current sensor and motion sensor are better than the other sensors. A multi sensor wireless system using current and motion sensor nodes have been fabricated and installed in the house of an elderly woman. It has been observed that the optimized system has been able to record faithfully the daily activities.