Wireless sensor networks (WSNs) have wide range of applications and provide limitless potential in our life. Unfortunately, they usually are prone to failure due to energy consumption, hardware failure, communication link errors, or malicious attacks. Therefore, fault tolerance mechanism is mandatory while designing the WSN. Fault tolerance includes fault detection, diagnosis, and repair. In the majority of WSN environments, fault tolerance is enforced and managed centrally at cluster head level. In this work, we present a distributed self-healing approach (DSHA), in which the processes of fault detection, diagnosis, and repair are performed at both node level and cluster head level. The proposed mechanism succeeded in locating hardware failures in sensor nodes, diagnosing them and applying countermeasures to ensure reliability and resiliency of the WSN. The countermeasures included isolating malfunction nodes and topology modifications. The experimental results proved that DSHA could tolerate up to 67.3% of hardware components failures and announce 62.6% improvement in the rate of sensor network lifetime.