Blind or visually impaired people feel miserable and helpless when people help them in their daily simple jobs. Many of them are afraid of going out unless it's necessary. Blind people usually memorize the location of objects so they can find them without struggling, but what if the location of any object has been changed for any reason? Or what if they want to go to an outside place? How they will recognize objects? Colors are also another challenge for them. Smart Assistant for blind people is a portable device. This device will make blind and visually impaired people's lives much easier, as it will help them in recognizing objects. Another aim is to identify texts on objects. An additional feature for the device is to make the device capable of recognizing different colors. This device is easy to use, and it can be used by users of different ages. There are two main parts for our project: software part and hardware part. Hardware part will contain camera, Ultrasonic sensor, Bluetooth headset, Raspberry Pi and power bank. While the software part will be OpenCV as the framework, that uses Python as the programming language. We also use a deep learning technique. This technique is a type of machine learning methods that uses neural network architectures to extract features from an image in order to recognize the object exists in the captured image.