Collaborative spectrum sensing improves the spectrum state estimation accuracy but is vulnerable to the potentialattacks from malicious secondary cognitive radio (CR) users,and thus raises security concerns. One promising malicioususer detection method is to identify their abnormal statisticalspectrum sensing behaviors. From this angle, two hidden Markovmodels (HMMs) corresponding to honest and malicious usersrespectively are adopted in this paper to characterize theirdifferent sensing behaviors, and malicious user detection isachieved via detecting the difference in the corresponding HMMparameters. To obtain the HMM estimates, an effective inferencealgorithm that can simultaneously estimate two HMMs withoutrequiring separated training sequences is also developed. Byusing these estimates, high malicious user detection accuracycan be achieved at the fusion center, leading to more robust andreliable collaborative spectrum sensing performance (substantially enlarged operational regions) in the presence of malicioususers, as compared to the baseline approaches. Different fusionmethods are also discussed and compared.