Wireless sensor systems, today widely adopted within the Internet of Things (IoT) paradigm, can collect and share large amount of data for different applications, but they are also vulnerable to cyberattacks. The impact of cyberattacks on the systems’ confidentiality, integrity and availability can be mitigated by the adoption of authentication procedures and cryptographic algorithms. Authentication passwords and cryptographic keys may be stored in a non-volatile memory, which may be easily tampered. Alternately, Physical Unclonable Functions (PUFs) can be adopted. They generate a chip’s unique fingerprint, by exploiting the randomness of process parameters’ variations occurring during chip fabrication, thus constituting a more secure alternate to the adoption of non-volatile memories for password storage. PUF reliability is of primary importance to guarantee system availability. In this paper, the reliability of a SRAM-based PUF implemented by a standard 32nm CMOS technology is investigated, as a function of different operating conditions, such as noise, power supply voltage and temperature, also considering different values of transistor conduction threshold voltages. The achieved results will show that transistor conduction threshold voltage tolerance and noise are the operating conditions that mostly affect PUF reliability, while the impact of temperature variations is lower, and the impact of power supply variations is negligible.