Mental health problems have increased to the extent that they have now become a social problem in many mature countries. It expected that patients recognize and control their condition themselves, through daily used devices and various countermeasures, and screening techniques are also useful for the patient. We have been conducting research on a smartphone application (mind monitoring system (MIMOSYS)), that estimates health conditions such as depression and stress state from the voice of the users. Vocal analysis using speech has the benefit of being easy to implement with minimal effort. In this research, to examine the effectiveness of the estimation of mental health condition from the voice by MIMOSYS in an iOS device, the analysis results of MIMOSYS was compared and its effectiveness was examined using the sound recorded by various iPhone devices. First, a healthy person recorded their voice by reading aloud fixed phrases. The recorded sound was then reproduced from the speaker and recorded as a control using a recorder (R-26), microphone (ME-52W), and iOS device. With the recorded voice at the microphone set as the control, the recorded voice at each iOS device and the MIMOSYS analysis result were compared and examined. As a result, a high correlation with the recording result and analysis with the microphone in any iPhone device was found. This confirms the useful smooth performance of mental health condition estimation by MIMOSYS speech on an iOS device.