Compared to wearable devices, handheld haptic devices are promising for large scale virtual reality applications because of their portability and capability of supporting large workspace haptic interaction. However, it remains a challenge to render multimodal haptic stimuli in handheld devices due to space confinement. In this paper, we present a modular approach to build a Multimodal Handheld Haptic Controller called “Haptic Handshank” that includes a thumb feedback component, a palm feedback component, and a motion tracking component. In the thumb feedback component, a compact pneumatically-driven silicone airbag is utilized to simulate softness, and a flexible membrane based on the electro-vibration principle which covers the top portion of the airbag for rendering virtual textures. In the palm feedback component, vibrational motors and Peltier devices are embedded into the device's body for rendering vibrotactile flow and distributing thermal stimuli. In the motion tracking component, an HTC-Vive tracker is mounted on the bottom of the controller's handle to enable 6-DOF palm motion tracking. The performance of the handheld device is evaluated through quantitative experimental studies, which validate the ability of the device to simulate multimodal haptic sensations in accordance with diverse hand manipulation gestures such as enclosure, static contact, rubbing, squeezing and shaking of a cup of cold drink in 3D virtual space.