With the continuous development of automotive intelligence, there is an increasing demand for vehicle chassis systems to become more intelligent, electronically controlled, integrated, and lightweight. In this context, the steer-by-wire system, which is electronically controlled, offers high precision and fast response. It provides greater flexibility, stability, and comfort for the vehicle, thus meeting the above requirements and has garnered widespread attention. Unlike traditional systems, the steer-by-wire system eliminates mechanical components, meaning the road feel cannot be directly transmitted to the steering wheel. To address this, the road feel, which is derived from the vehicle's state or integrated with environmental driving data, must be simulated and transmitted to the steering wheel through a road feel motor. This motor generates feedback that mimics the road feel, similar to that experienced in a conventional steering system. This simulation enhances the driver's experience and can improve driving safety. Therefore, studying road feel simulation in steer-by-wire systems is of significant importance. The road feel simulation motor plays a crucial role in a steer-by-wire system. Non-direct drive motors have limited torque and generally require the addition of devices such as mechanisms to change the transmission ratio in order to increase torque. This can introduce gaps, mechanical losses, noise, and additional friction resistance, ultimately reducing transmission efficiency. However, these components can introduce gaps, mechanical losses, noise, and additional friction, leading to reduced transmission efficiency. On the other hand, direct drive motors eliminate many of these components and transmission losses, offering higher transmission efficiency, faster response, and greater accuracy. However, direct drive motors tend to be more expensive. This paper compares various dynamic model estimation and parameter fitting methods for road feel simulation in steer-by-wire systems. It also designs a road feel simulation algorithm based on a dynamic model. The performance of the two types of motors is evaluated through hardware-in-the-loop (HIL) testing, using an NI PXI industrial computer as the hardware controller and NI VeriStand for hardware-software signal transmission on the test bench.