This study investigated the factors influencing the measurement of bubble velocity in gas-liquid two-phase flow using the ultrasonic echo method, focusing on a typical gas slug flow pattern. A theoretical model was developed using pulse random signals with a certain time delay as input for the upstream and downstream sensors. The effects of time delay, noise, and pulse signals on the cross-correlation analysis of the random signals were analyzed. The results show that the cross-correlation coefficient decreases as the time delay increases. After adding white noise, the cross-correlation coefficient increases with signal-to-noise ratio, though the rate of increase gradually slows down. With the addition of pulse signals, the coefficient increases as the number, height, and width of the pulses increase, with the height of the pulses having a greater impact on the cross-correlation coefficient than the width. Based on this, an experimental platform for gas-liquid two-phase flow was established to investigate the effects of sensor spacing, gas flow rate, data analysis length, and ultrasonic transducer gain on the cross-correlation analysis of ultrasonic echo signals. The results show that as the sensor spacing increases, the bubble velocity error increases and the cross-correlation coefficient decreases. As the gas flow rate increases, the bubble velocity error reaches a minimum at 3 L/min, while the cross-correlation coefficient increases. Increasing the data analysis length results in a velocity error that remains within 10% when the number of bubbles exceeds 4, with the overall cross-correlation coefficient showing an upward trend. As the ultrasonic transducer gain increases, the cross-correlation coefficient consistently shows strong correlation, with errors staying within 10% only when the gain is between 28 dB and 32 dB. This provides valuable insights for the development, system design, optimization, and industrial application of ultrasonic echo cross-correlation methods for bubble velocity measurement.