OBJECTIVE：To study the effect of paroxetine combined with low dose of olanzapine on sleep process and architecture of depression patients with insomnia. METHODS： 84 depression patients with insomnia were randomly divided into control group and observation group. Control group was given 20 mg Paroxetine tablet， once every morning； observation group was additionally given 2.5 mg Olanzapine tablet， once before going to bed. Sleep quality [Pittsburgh Sleep Quality Index （PSQI）]， depression scores [Hamilton Depression Scale （HAMD）]， sleep process [sleep latency （SL）， awakening times （AT）， the actual total sleep time （TST）， sleep efficiency （SE）， rapid eye movement （REM） sleep latency （RL）] and sleep architecture [sleep stage 1 （S1），2（S2），3（S3） and the proportion of REM to sleep] in 2 groups before and 3， 6 months after treatment and the incidence of adverse reactions（ADR） were observed. RESULTS： After treatment， PSQI and HAMD scores in 2 groups were significantly lower than before， and gradually decreased by time， and observation group was lower than control group； TST in observation group was significantly higher than before and control group， S1 in observation group was significantly lower than before， SE， S3 and REM in 2 groups were significantly higher than before， and observation group was higher than control group， SL， AT， RL and S2 were significantly lower than before， and observation group was lower than control group， the differences were statistically significant（P＜0.05）. There were no obvious ADR in 2 groups. CONCLUSIONS： Paroxetine combined with low dose of olanzapine can significantly relieve depression， optimize sleep process and sleep architecture， then impove sleep quality.