The 5G millimetre wave communication has attracted much attention due to its high carrier frequency and rate. It is a hot research topic for future mobile communications. However, in the design of millimetre-wave communication test instruments, multiple frequency multiplication is usually used to provide millimetre-wave fundamental oscillation for millimetre-wave base wave mixers, and the errors such as phase noise, carrier frequency offset and sampling clock offset additionally generated in this process can cause deterioration of the system demodulation index. To solve this problem, this article presents for the first time a method for comprehensive estimation of millimetre-wave errors based on the particle filtering algorithm, which can estimate the phase noise, carrier frequency offset and sampling clock offset in one go using the good robustness and adaptability of the particle filtering algorithm. Through simulation and experimental analysis, this article shows that the algorithm is more effective than the traditional PTRS phase noise estimation, and the EVM index is better. The proposed algorithm has been applied to the Chinese 5G millimetre wave base station synthesizer, and the EVM index can reach 2. 21% at a carrier frequency of 28 GHz, a bandwidth of 400 MHz and a modulation method of 64QAM.