A handheld device, an object positioning method thereof and a computer-readable recording medium are provided. The handheld device includes a radar sensor, an image sensor and a control unit. The radar sensor emits a detection wave, and receives a reflected wave generated by an object by reflecting the detection wave. Each object generates one of the reflected waves. The image sensor captures an image. The image includes a subset of the objects. The control unit extracts a waveform signature of each reflected wave, recognizes the waveform signature in a time domain and a frequency domain to determine a first type of each object, obtains a first position of the object according to the reflected wave, obtains a second type and a second position of each object according to the first image, and performs object mapping to combine or compare the first position and the second position of the object.