diff --git a/README.md b/README.md
index 1981257..c384f5f 100644
--- a/README.md
+++ b/README.md
@@ -55,24 +55,45 @@ git submodule init
 
 git submodule update
 
-xmake b crossdesk
+xmake b -vy crossdesk
 ```
-#### 无 CUDA 环境下的开发支持
 
-对于未安装 **CUDA 环境** 的Linux开发者，这里提供了预配置的 [Ubuntu 22.04 Docker 镜像](https://hub.docker.com/r/crossdesk/ubuntu22.04)。  
-该镜像内置必要的构建依赖，可在容器中开箱即用，无需额外配置即可直接编译项目。
+运行
+```
+xmake r crossdesk
+```
+
+### 无 CUDA 环境下的开发支持
+
+对于未安装 **CUDA 环境** 的 Linux 开发者，这里提供了预配置的 [Ubuntu 22.04 Docker 镜像](https://hub.docker.com/r/crossdesk/ubuntu22.04)。该镜像内置必要的构建依赖，可在容器中开箱即用，无需额外配置即可直接编译项目。
 
 进入容器，下载工程后执行：
 ```
 export CUDA_PATH=/usr/local/cuda
 export XMAKE_GLOBALDIR=/data
 
-xmake b --root crossdesk
+xmake b --root -vy crossdesk
 ```
 
-运行
+对于未安装 **CUDA 环境** 的 Windows 开发者，执行下面的命令安装 CUDA 编译环境：
 ```
-xmake r crossdesk
+xmake require -vy "cuda 12.6.3"
+```
+安装完成后执行:
+```
+xmake require --info "cuda 12.6.3"
+```
+输出如下:
+
+<img width="860" height="226" alt="Image" src="https://github.com/user-attachments/assets/999ac365-581a-4b9a-806e-05eb3e4cf44d" />
+
+根据上述输出获取到 CUDA 的安装目录，即 installdir 指向的位置。将 CUDA_PATH 加入系统环境变量，或在终端中输入：
+```
+set CUDA_PATH=path_to_cuda_installdir
+```
+重新执行：
+```
+xmake b -vy crossdesk
 ```
 
 #### 注意
diff --git a/README_EN.md b/README_EN.md
index fef063f..6889f5d 100644
--- a/README_EN.md
+++ b/README_EN.md
@@ -55,12 +55,17 @@ git submodule init
 
 git submodule update
 
-xmake b crossdesk
+xmake b -vy crossdesk
+```
+
+Run:
+```
+xmake r crossdesk
 ```
 
 #### Development Without CUDA Environment
 
-For developers who do not have a **CUDA environment** installed, a preconfigured [Ubuntu 22.04 Docker image](https://hub.docker.com/r/crossdesk/ubuntu22.04) is provided.  
+For **Linux** developers who do not have a **CUDA environment** installed, a preconfigured [Ubuntu 22.04 Docker image](https://hub.docker.com/r/crossdesk/ubuntu22.04) is provided.  
 This image comes with all required build dependencies and allows you to build the project directly inside the container without any additional setup.
 
 After entering the container, download the project and run:
@@ -68,12 +73,29 @@ After entering the container, download the project and run:
 export CUDA_PATH=/usr/local/cuda
 export XMAKE_GLOBALDIR=/data
 
-xmake b --root crossdesk
+xmake b --root -vy crossdesk
 ```
 
-Run:
+For **Windows** developers without a **CUDA environment** installed, run the following command to install the CUDA build environment:
 ```
-xmake r crossdesk
+xmake require -vy "cuda 12.6.3"
+```
+After the installation is complete, execute:
+```
+xmake require --info "cuda 12.6.3"
+```
+The output will look like this:
+
+<img width="860" height="226" alt="Image" src="https://github.com/user-attachments/assets/999ac365-581a-4b9a-806e-05eb3e4cf44d" />
+
+From the output above, locate the CUDA installation directory — this is the path pointed to by installdir.
+Add this path to your system environment variable CUDA_PATH, or set it in the terminal using:
+```
+set CUDA_PATH=path_to_cuda_installdir:
+```
+Then re-run:
+```
+xmake b -vy crossdesk
 ```
 
 #### Notice